arXiv:2112.00789v1 [astro-ph.GA] 1 Dec 2021
Transcript of arXiv:2112.00789v1 [astro-ph.GA] 1 Dec 2021
Draft version December 3, 2021Typeset using LATEX twocolumn style in AASTeX63
Galactic Winds across the Gas-Rich Merger Sequence.I. Highly Ionized N V and O VI Outflows in the QUEST Quasars∗
Sylvain Veilleux,1, 2 David S. N. Rupke,3 Weizhe Liu,1 Anthony To,3, 4 Margaret Trippe,1, 5, 6
Todd M. Tripp,7 Fred Hamann,8 Reinhard Genzel,9 Dieter Lutz,9 Roberto Maiolino,10, 11 Hagai Netzer,12
Kenneth R. Sembach,13 Eckhard Sturm,9 Linda Tacconi,9 and Stacy H. Teng1, 14
1Department of Astronomy, University of Maryland, College Park, MD 20742, USA2Joint Space-Science Institute, University of Maryland, College Park, MD 20742, USA
3Department of Physics, Rhodes College, Memphis, TN 38112, USA4Department of Physics, University of Hawaii, Honolulu, HI 96822, USA
5Johns Hopkins University Applied Physics Laboratory, Laurel, MD 20723 USA6Lincoln Laboratory, Massachusetts Institute of Technology, Lexington, MA 02421-6426 USA
7Department of Astronomy, University of Massachussetts, Amherst, MA 01003, USA8Department of Physics and Astronomy, University of California, Riverside, CA 92507, USA
9Max-Planck-Institut für Extraterrestrische Physik, Giessenbachstrasse 1, 85748 Garching, Germany10Cavendish Laboratory, University of Cambridge, 19 J.J. Thomson Avenue, Cambridge CB3 0HE, United Kingdom11Kavli Institute for Cosmology, University of Cambridge, Madingley Road, Cambridge CB3 0HA, United Kingdom
12School of Physics and Astronomy, Tel-Aviv University, Tel Aviv 69978, Israel13Space Telescope Science Institute, Baltimore, MD 21218, USA14Institute for Defense Analyses, Alexandria, MD 22311, USA
ABSTRACTThis program is part of QUEST (Quasar/ULIRG Evolutionary Study) and seeks to examine the
gaseous environments of z . 0.3 quasars and ULIRGs as a function of host galaxy properties andage across the merger sequence from ULIRGs to quasars. This first paper in the series focuses on33 quasars from the QUEST sample and on the kinematics of the highly ionized gas phase traced bythe N V λλ1238,1243 and O VI λλ1032,1038 absorption lines in high-quality Hubble Space Telescope(HST) Cosmic Origins Spectrograph (COS) data. N V and O VI outflows are present in about 60% ofthe QUEST quasars and span a broad range of properties, both in terms of equivalent widths (from 20mÅ to 25 Å) and kinematics (outflow velocities from a few × 100 km s−1 up to ∼10,000 km s−1). Therate of incidence and equivalent widths of the highly ionized outflows are higher among X-ray weakor absorbed sources. The weighted outflow velocity dispersions are highest among the X-ray weakestsources. No significant trends are found between the weighted outflow velocities and the properties ofthe quasars and host galaxies although this may be due to the limited dynamic range of propertiesof the current sample. These results will be re-examined in an upcoming paper where the sampleis expanded to include the QUEST ULIRGs. Finally, a lower limit of ∼0.1% on the ratio of time-averaged kinetic power to bolometric luminosity is estimated in the 2−4 objects with blueshifted P Vλλ1117,1128 absorption features.
Keywords: galaxies: evolution − ISM: jets and outflows − quasars: absorption lines − quasars: general
Corresponding author: Sylvain [email protected]
∗ Based on observations made with the NASA/ESA Hubble SpaceTelescope, obtained from the data archive at the Space TelescopeScience Institute. STScI is operated by the Association of Uni-versities for Research in Astronomy, Inc. under NASA contractNAS 5-26555.
1. INTRODUCTION
Large multiwavelength surveys of the local and dis-tant universe have shown that major mergers of gas-
arX
iv:2
112.
0078
9v1
[as
tro-
ph.G
A]
1 D
ec 2
021
2 Veilleux et al.
rich galaxies1 may trigger spectacular bursts of star for-mation, accompanied with quasar-like episodes of rapidgrowth of the supermassive black holes (SMBHs), andresult in merger remnants that follow tight SMBH-hostscaling relations and resemble today’s quiescent early-type galaxies (e.g. Sanders et al. 1988; Hickox & Alexan-der 2018). Modern simulations of galaxy formation andevolution (e.g. Nelson et al. 2019; Oppenheimer et al.2020; Nelson et al. 2021) largely reproduce these obser-vations. However, the root cause of the fast “quenching”of the star formation activity in the merger remnantsdepends on the detailed, sub-grid scale, implementationof how the mass, momentum, and energy from stellarwinds, supernova explosions, and SMBH-related pro-cesses are injected into, and interact with, the interstel-lar medium (ISM) and circumgalactic medium (CGM)of the host galaxies. Over the past several years, nearbygas-rich galaxy mergers have emerged as excellent labo-ratories to study in detail these stellar and quasar feed-back processes (see Veilleux et al. 2020, for a recent re-view). These objects are the focus of the present study.Locally, major gas-rich galaxy mergers often co-
incide with obscured ultraluminous infrared galaxies(ULIRGs). As these systems evolve, the obscuring gasand dust, funneled to the center by the dissipative col-lapse and tidal forces during the merger, are either trans-formed into stars or expelled out of the nucleus by pow-erful winds driven by the central quasar and starburst,giving rise to dusty quasars and finally to completelyexposed quasars. Galactic-scale winds are ubiquitous inlocal ULIRGs and dusty quasars (e.g. Sturm et al. 2011;Veilleux et al. 2013a; Cicone et al. 2014; Rupke et al.2017; Veilleux et al. 2017; Fluetsch et al. 2019, 2020;Lutz et al. 2020; Veilleux et al. 2020, and referencestherein). The outflows detected in ULIRGs extend overa large range of distances from the central energy source,seamlessly blending with the circumgalactic medium at>10 kpc (Veilleux et al. 2020, and references therein).In these objects, the outflow masses and energetics are
often dominated by the outer (&kpc) cool dusty molec-ular or neutral atomic gas phase, but the driving mech-anism is best probed by examining the inner (. sub-kpc) ionized phase. ULIRG F11119+3257 is the firstand still the best case among local ULIRGs where a fast(>0.1 c), highly ionized (Fe XXV/XXVI at ∼7 keV),accretion-disk scale (<1 pc) quasar wind appears to bedriving a massive (> 100 M yr−1), large-scale (1−10+kpc) molecular and neutral-gas outflow (Tombesi et al.
1 In this paper, we define major mergers as those involving galaxieswith . 4:1 stellar mass ratios.
2015, 2017; Veilleux et al. 2017). Unfortunately, thesearch for hot winds in a statistically significant sampleof ULIRGs is not feasible at present since most ULIRGsare too faint at ∼7 keV for current X-ray observatories.This is where the excellent far-ultraviolet (FUV)
spectroscopic sensitivity of the Hubble Space Telescope(HST) becomes handy. The FUV band is rich in spec-troscopic diagnostics of the neutral, low-ionization, andhigh-ionization gas phases (Haislmaier et al. 2021); thusHST can probe all three phases at once. So far, onlyabout a dozen ULIRGs and IR-bright quasars havebeen studied with HST, but the results have beenpromising. Prominent, blueshifted Lyα emission outto −1000 km s−1 has been detected in half of theseULIRGs. Blueshifted absorption features from high-ionization species like N V and O VI (77 and 114 eVare needed to produce N4+ and O5+ ions, respectively)and/or low-ionization species like Si II, Si III, Fe II,N II, and Ar I have provided additional unambiguoussignatures of outflows in a few of these objects. Martinet al. (2015) have argued that the FUV-detected out-flows represent clumps of gas condensing out of a fast,hot wind generated by the central starburst (Thomp-son et al. 2016). This picture is also consistent with theblast-wave model for quasar feedback. In this model, afast, hot wind shocks the surrounding ISM, which theneventually cools to reform the molecular gas after havingacquired a significant fraction of the initial kinetic en-ergy of the hot wind (e.g. Weymann et al. 1985; Zubovas& King 2012, 2014; Faucher-Giguère & Quataert 2012;Nims et al. 2015; Richings & Faucher-Giguère 2018a,b;Richings et al. 2021; Girichidis et al. 2021). An alterna-tive explanation is radiative acceleration (e.g. Ishibashiet al. 2018, 2021), which may dominate the dynamicsof outflows on a wide variety of scales (e.g. Stern et al.2016; Revalski et al. 2018; Somalwar et al. 2020).So far, the published data set on ULIRGs and
IR-bright quasars is too small to draw strong con-clusions about the properties of the FUV-detectedwinds. There is tantalizing evidence that UV-detected AGN/starburst-driven winds are present inmost ULIRGs, but the sample is very incomplete, partic-ularly among ULIRGs with AGN and matched quasars.A more diverse sample of ULIRGs and quasars is neededto study the gaseous environments of nearby quasars andULIRGs as a function of host properties and age acrossthe merger sequence from ULIRGs to quasars. This is-sue is addressed in the present study.In this first paper, we focus our efforts on studying
the highly ionized gas, traced by N V λλ1238, 1243 andO VI λλ1032, 1038, in a sample of 33 local quasars,while Paper II (Liu et al. 2021, in prep.) will present
Highly Ionized N V and O VI Outflows in the QUEST Quasars 3
the results on our sample of ULIRGs with AGN andcompare them with those on the quasars. As statedin Hamann et al. (2019b), the quasars in the presentsample are valuable for outflow studies in and of them-selves because: 1) they fill a largely-unexplored nichebetween luminous quasars with strong broad absorptionlines (BALs) with outflow velocities of up to 0.1 − 0.2c and low-luminosity Seyfert 1 galaxies with exclusivelynarrower outflow lines, 2) their low redshift minimizescontamination by the Lyα forest, and 3) the outflowlines are relatively narrow so blending is less severe. In-deed, as we discuss below, the detected outflows oftenare “mini-BALs” instead of BALs because their velocitywidths lie below or near the threshold of 2000 km s−1
used for BALs (Weymann et al. 1981, 1991; Hamann &Sabra 2004; Gibson et al. 2009b).Our quasar sample is discussed in Section 2. The ex-
tensive set of ancillary data on these quasars is summa-rized in Section 3. The HST spectra used for this studyare described in Section 4 and the methods applied toanalyze these data are detailed in Section 5. The resultsfrom this analysis are presented in Section 6, and dis-cussed in more detail in Section 7. Section 8 provides asummary of the main results from this paper.
2. QUASAR SAMPLE
The quasars in our sample are selected using fourcriteria: (1) They must be part of the QUEST(Quasar/ULIRG Evolutionary Study) sample of local(z . 0.3) ULIRGs and quasars. The QUEST samplehas already been described in detail in Veilleux et al.(2009a,b) and references therein. All 33 objects in thepresent sample are Palomar-Green (PG) quasars fromthe Bright Quasar Sample (Schmidt & Green 1983), ex-cept Mrk 231, the nearest quasar known, whose UVspectrum has already been analyzed by Veilleux et al.(2013b, 2016) and will not be discussed here any further.As part of the QUEST sample, the quasars are carefullymatched in terms of redshifts, bolometric luminosities,and host galaxy masses with the QUEST ULIRGs of Pa-per II. (2) Their bolometric luminosity must be quasar-like, & 1045 ergs s−1, and dominated by the quasarrather than the starburst based on the Spitzer data (seecriterion #3 below) or, equivalently, have 25-to-60 µmflux ratios f25/f60 & 0.15 (Veilleux et al. 2009a). Thiscriterion also automatically selects UV-detected late-stage mergers or non-mergers (Veilleux et al. 2009b). (3)A strong preference is also given to the QUEST quasarswith Spitzer mid-infrared spectra to provide valuable in-formation on the AGN contribution to the bolometric lu-minosities of these objects. (4) High-quality COS spec-tra covering systemic N V and/or O VI must exist for
each object in the sample. Only COS data are consid-ered to ease comparisons between spectra and avoid pos-sible systematic errors associated with comparing datasets from different instruments. As described in Sec-tion 4, both our own and archival data are used for thisstudy.These criteria result in a sample of 33 objects. Table
1 lists the key properties of the quasars in our sam-ple, many of which are derived from our extensive setof ancillary data on these objects, discussed in Section3. As shown in Figure 1, these quasars cover the low-redshift and low bolometric luminosity ends of the PGquasar sample. They are well matched in redshift withthe QUEST ULIRGs which will be studied in Paper II(Liu et al. 2021, in prep), and are representative of theentire PG quasars sample in terms of infrared excess(defined here as the infrared-to-bolometric luminosityratio, LIR/LBOL) and FIR brightness (L60 µm/L15 µm
from Netzer et al. 2007).
3. ANCILLARY DATA
An extensive set of spectroscopic and photometricdata exist on all of the objects in the sample. SloanDigital Sky Survey (SDSS) optical spectra are availablefor all of them. High-quality optical spectra also existin Boroson & Green (1992), and Krug (2013) presentsspectra centered on Na I D λλ5890, 5896. As mentionedin Section 2, most of these quasars have also been stud-ied spectroscopically in the mid-infrared with Spitzer(Schweitzer et al. 2006, 2008; Netzer et al. 2007; Veilleuxet al. 2009a). In addition, nearly all of the quasars inthis sample are part of X-QUEST, an archival XMM-Newton and Chandra X-ray spectroscopic survey of theQUEST sample (Teng & Veilleux 2010; Columns 13-17in Table 1). VLT and Keck near-infrared spectroscopicdata exist for a number of these objects (Dasyra et al.2007).Optical and near-infrared images of these objects have
been obtained from the ground (Surace et al. 2001;Veilleux et al. 2002; Guyon et al. 2006) and with HST(Veilleux et al. 2006; Kim et al. 2008; Hamilton et al.2008; Veilleux et al. 2009b), providing photometric andmorphological measurements on both the quasars andhost galaxies (e.g., morphological type, quasar-to-hostluminosity ratio, strength of tidal features). Far-infraredphotometry obtained with the Herschel PACS instru-ment exists for all of these objects (Lani et al. 2017;Shangguan et al. 2018), while far-infrared spectra cen-tered on the OH 119 µm feature exist for five of them(Veilleux et al. 2013a). Finally, Green Bank Telescope(GBT) H I 21-cm line emission and absorption spectraare available for 16 of these quasars (Teng et al. 2013).
4 Veilleux et al.
The connection between UV and X-ray propertiesis critical, so we searched the literature for addi-tional X-ray measurements (ignoring older ones fromROSAT/ASCA). These are listed and referenced in Ta-ble 1. For Chandra observations from Teng & Veilleux
(2010) where no absorbing column was detected, morerecent constraints from Ricci et al. (2017) (based onSwift/BAT detections) are available in some cases. Inthese cases, we substitute the newer measurement of ab-sorbing column.
Highly Ionized N V and O VI Outflows in the QUEST Quasars 5
Tab
le1.
Prope
rtiesof
theQUEST
Qua
sars
intheSa
mple
Nam
eOther
zlogνLν(U
V)logR
Rad
ioα
OX
log(L
bol
L
)α
AG
Nlog(L
IRL
)log(M
BH
M
)logη
Edd
log(L
SX)
log(L
HX)
ΓX
NH
Ref.
Nam
e[erg
s−1]
Class
[erg
s−1]
[erg
s−1]
[102
2cm
−2]
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
(15)
(16)
(17)
PG
0007
+10
6IIIZw
20.08
9344
.55
+2.29
Flat−1.43
12.24
1a11
.63
8.07
+0.4
5−
0.4
6−0.34
+0.4
6−
0.4
543
.94
44.18
1.73
+0.0
4−
0.0
40.11
0.0
2−
0.0
1a,
4,5
PG
0026
+12
9...
0.14
5445
.32
+0.03
Quiet−1.50
12.08
0.98
6+0.0
14
−0.0
27
11.71
8.49
+0.4
4−
0.4
5−0.93
+0.4
5−
0.4
444
.40
2.00
+0.1
3−
0.1
1<0.01
g,5
PG
0050
+12
4IZw
10.05
8944
.24
−0.48
Quiet−1.56
12.08
0.92
5+0.0
75
−0.0
94
12.04
7.33
+0.6
2−
0.6
20.20
+0.6
2−
0.6
244
.04+
0.0
5−
0.0
943
.88+
0.0
5−
0.0
92.25
+0.0
5−
0.0
30.09
+0.0
8−
0.0
2f,6
43.78+
0.0
2−
0.1
243
.64+
0.0
1−
0.0
42.09
+0.0
3−
0.0
30.04
+0.0
3−
0.0
26
43.633
+0.0
05
−0.0
052.37
+0.0
8−
0.0
40.04
53
PG
0157
+00
1Mrk
1014
0.16
3345
.34
+0.33
Quiet−1.60
12.70
0.72
7+0.2
36
−0.2
71
12.67
8.06
+0.6
1−
0.6
2−0.01
+0.6
3−
0.6
543
.92+
0.0
4−
0.0
443
.83+
0.1
1−
0.2
52.54
+0.0
9−
0.0
9<0.00
9e,
3,6
44.00+
0.0
4−
0.0
843
.86+
0.0
4−
0.0
42.1+
0.1
−0.1
<0.00
93,
6
PG
0804
+76
1...
0.10
045
.54
−0.22
Quiet−1.52
12.09
0.99
6+0.0
04
−0.0
04
11.98
8.73
+0.4
3−
0.4
3−1.16
+0.4
3−
0.4
344
.54
44.45
2.27
+0.0
9−
0.2
00.04
4+0.0
07
−0.0
10
a
PG
0838
+77
0VII
Zw
244
0.13
2444
.83
−0.96
Quiet−1.54
11.77
0.94
5+0.0
27
−0.0
18
11.66
8.05
+0.6
1−
0.6
2−0.82
+0.6
2−
0.6
144
.152
+0.0
09
−0.0
14
43.54+
0.0
4−
0.0
51.49
+0.0
8−
0.0
8<0.1b
d,4,
5
PG
0844
+34
9...
0.06
444
.59
−1.52
Quiet−1.54
11.45
0.97
1+0.0
29
−0.0
49
11.18
7.86
+0.4
6−
0.4
9−0.94
+0.4
9−
0.4
644
.152
+0.0
09
−0.0
14
43.80+
0.0
3−
0.0
32.66
+0.0
5−
0.0
66.13
+3.0
3−
1.3
9a,
6
PG
0923
+20
1...
0.19
245
.42
−0.85
Quiet−1.57
12.46
0.99
0+0.0
00
−0.0
00
12.05
7.90
+0.6
1−
0.6
20.04
+0.6
2−
0.6
1i
PG
0953
+41
4...
0.23
4145
.95
−0.36
Quiet−1.50
12.53
0.98
2+0.0
18
−0.0
18
12.20
8.33
+0.4
4−
0.4
4−0.33
+0.4
4−
0.4
445
.037
+0.0
06
−0.0
09
44.81+
0.0
2−
0.0
32.44
+0.0
3−
0.0
318
.52+
9.8
4−
5.6
h,6
PG
1001
+05
4...
0.16
1144
.93
−0.30
Quiet−2.13
11.87
0.83
6+0.1
23
−0.1
00
11.66
7.63
+0.6
1−
0.6
2−0.35
+0.6
2−
0.6
243
.00+
0.1
8−
0.3
043
.08+
0.1
5−
0.6
02.01
+0.6
7−
0.4
88.09
+5.4
7−
3.5
7e,
6
PG
1004
+13
04C
+13
.41
0.24
0645
.30
+2.36
Steep<−2.01
12.69
0.96
3+0.0
37
−0.0
36
12.22
9.16
+0.6
1−
0.6
2−1.01
+0.6
2−
0.6
143
.48+
0.0
5−
0.0
543
.76+
0.0
8−
0.1
31.67
+0.2
0−
0.1
12.99
+2.6
7−
1.3
7e,
6
43.51+
0.0
4−
0.1
543
.89+
0.0
7−
0.2
21.52
+0.1
7−
0.2
61.44
+0.6
4−
0.6
96
PG
1116
+21
5...
0.17
6545
.79
−0.14
Quiet−1.57
12.55
0.99
1+0.0
09
−0.0
08
12.28
8.42
+0.6
1−
0.6
2−0.39
+0.6
2−
0.6
144
.927
+0.0
21
−0.0
06
44.65+
0.0
3−
0.0
42.53
+0.0
4−
0.0
327
.21+
16.0
1−
11.2
6h,
6
44.922
+0.0
04
−0.0
04
44.65+
0.0
3−
0.0
42.49
+0.0
1−
0.0
131
.61+
5.1
4−
4.1
36
44.93+
0.0
1−
0.0
144
.67+
0.0
3−
0.0
32.51
+0.0
4−
0.0
420
.21+
5.9
4−
5.1
06
PG
1126−04
1Mrk
1298
0.06
044
.29
−0.77
Quiet−2.13
11.53
0.96
2+0.0
38
−0.0
75
11.52
7.64
+0.6
1−
0.6
2−0.65
+0.6
2−
0.6
143
.04+
0.0
5−
0.0
543
.11+
0.0
5−
0.1
21.95
+0.1
0−
0.1
04.66
+0.4
2−
0.3
9a,
6
PG
1211
+14
3...
0.08
0944
.96
+1.39
Steep−1.57
11.97
1.00
0+0.0
00
−0.0
00
11.74
7.85
+0.6
1−
0.6
2−0.40
+0.6
2−
0.6
144
.328
+0.0
05
−0.0
07
43.94+
0.0
1−
0.0
12.83
+0.0
2−
0.0
212
.98+
0.9
4−
0.9
0h,
6
44.201
+0.0
05
−0.0
05
43.89+
0.0
1−
0.0
12.63
+0.0
2−
0.0
212
.40+
1.6
4−
1.4
96
PG
1226
+02
33C
273
0.15
846
.50
+3.06
Flat−1.47
13.03
0.94
9+0.0
51
−0.1
28
12.80
8.41
+0.1
5−
0.2
40.08
+0.2
4−
0.1
645
.491
+0.0
02
−0.0
0245
.742
+0.0
08
−0.0
082.07
+0.0
1−
0.0
1<0.01
a,5,
6
45.461
+0.0
04
−0.0
0445
.722
+0.0
05
−0.0
061.81
+0.0
1−
0.0
1<0.01
5,6
45.591
+0.0
02
−0.0
0245
.820
+0.0
04
−0.0
052.28
+0.0
1−
0.0
1<0.01
5,6
45.663
+0.0
02
−0.0
0145
.825
+0.0
06
−0.0
062.08
+0.0
1−
0.0
1<0.01
5,6
45.461
+0.0
03
−0.0
04
45.67+
0.0
1−
0.0
22.13
+0.0
2−
0.0
2<0.01
5,6
45.544
+0.0
03
−0.0
0345
.941
+0.0
10
−0.0
071.96
+0.0
1−
0.0
2<0.01
5,6
PG
1229
+20
4Mrk
771
0.06
444
.42
−0.96
Quiet−1.49
11.57
0.98
5+0.0
15
−0.0
30
11.27
7.76
+0.4
6−
0.4
8−0.71
+0.4
8−
0.4
643
.785
+0.0
07
−0.0
08
43.61+
0.0
2−
0.0
22.38
+0.0
3−
0.0
313
.52+
5.7
7−
3.3
6g,
6
Mrk
231
...
0.04
217
42.70
−1.92
12.61
0.70
9+0.0
66
−0.0
67
12.54
8.58
+0.5
0−
0.5
0−0.63
+0.5
0−
0.5
042
.13+
0.0
1−
0.0
442
.58+
0.0
1−
0.1
11.40
+0.0
3−
0.1
9.5+
2.3
−1.9
b,7,
8
19.4
+5.7
−4.4
7,8
PG
1302−10
2PKS13
02-102
0.27
8445
.83
+2.27
Flat−1.58
12.75
0.98
2+0.0
18
−0.0
37
12.49
8.77
+0.6
1−
0.6
2−0.55
+0.6
2−
0.6
144
.81
1.66
+0.1
0−
0.1
1<0.06
h,1
PG
1307
+08
5...
0.15
545
.35
−1.00
Quiet−1.52
12.35
0.95
2+0.0
48
−0.0
66
11.76
8.54
+0.4
4−
0.4
6−0.72
+0.4
6−
0.4
444
.02+
0.0
2−
0.0
244
.16+
0.0
5−
0.0
91.89
+0.1
1−
0.1
05.64
+2.6
2−
1.4
8a,
6
PG
1309
+35
5...
0.18
2945
.05
+1.26
Flat−1.71
12.32
0.87
0+0.1
30
−0.1
27
12.05
8.24
+0.6
1−
0.6
2−0.49
+0.6
2−
0.6
243
.87+
0.0
2−
0.0
143
.88+
0.0
4−
0.0
52.19
+0.0
7−
0.0
66.02
+3.6
8−
1.8
4i,6
PG
1351
+64
0...
0.08
8245
.22
+0.64
Quiet−1.78
12.05
0.77
9+0.1
43
−0.2
21
11.87
8.72
+0.6
1−
0.6
2−1.30
+0.6
2−
0.6
343
.398
+0.0
07
−0.0
23
43.23+
0.0
3−
0.0
42.42
+0.0
4−
0.0
414
.61+
5.7
2−
3.8
1h,
6
Tab
le1continued
6 Veilleux et al.Tab
le1(con
tinu
ed)
Nam
eOther
zlogνLν(U
V)logR
Rad
ioα
OX
log(L
bol
L
)α
AG
Nlog(L
IRL
)log(M
BH
M
)logη
Edd
log(L
SX)
log(L
HX)
ΓX
NH
Ref.
Nam
e[erg
s−1]
Class
[erg
s−1]
[erg
s−1]
[102
2cm
−2]
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
(15)
(16)
(17)
PG
1411
+44
2...
0.08
9644
.34
−0.89
Quiet−2.03
11.79
1.00
0+0.0
00
−0.0
00
11.66
8.54
+0.4
5−
0.4
6−1.27
+0.4
6−
0.4
543
.60+
0.0
5−
0.0
643
.41+
0.0
6−
0.1
82.41
+0.1
8−
0.1
526
.29+
3.7
6−
4.0
8h,
6
PG
1435−06
7...
0.12
945
.12
−1.15
Quiet−1.63
11.92
0.97
6+0.0
24
−0.0
40
11.47
8.26
+0.6
1−
0.6
2−0.86
+0.6
2−
0.6
144
.11+
0.0
4−
0.0
443
.94+
0.0
7−
0.0
82.36
+0.1
1−
0.1
0<0.1b
a,6
PG
1440
+35
6Mrk
478
0.07
745
.18
−0.43
Quiet−1.38
11.81
0.83
6+0.0
71
−0.0
81
11.76
7.36
+0.6
1−
0.6
2−0.15
+0.6
2−
0.6
144
.33+
0.0
1−
0.0
143
.74+
0.0
5−
0.0
63.02
+0.0
4−
0.0
48.92
+8.6
6−
3.5
1a,
6
44.375
+0.0
05
−0.0
04
43.90+
0.0
1−
0.0
12.86
+0.0
1−
0.0
114
.24+
2.8
8−
2.3
46
44.299
+0.0
06
−0.0
06
43.74+
0.0
3−
0.0
32.98
+0.0
2−
0.0
28.22
5+2.0
3−
1.4
96
44.127
+0.0
07
−0.0
08
43.64+
0.0
2−
0.0
32.86
+0.0
2−
0.0
212
.67+
3.4
5−
2.7
06
PG
1448
+27
3...
0.06
543
.78
−0.60
Quiet−1.59
11.44
0.99
7+0.0
03
−0.0
07
11.19
6.86
+0.6
1−
0.6
20.06
+0.6
2−
0.6
143
.949
+0.0
06
−0.0
07
43.49+
0.0
2−
0.0
22.80
+0.0
2−
0.0
116
.72+
6.2
4−
4.3
9a,
6
PG
1501
+10
6Mrk
841
0.03
644
.03
−0.44
Quiet−1.64
11.34
1.00
0+0.0
00
−0.0
00
11.13
8.42
+0.6
1−
0.6
2−1.59
+0.6
2−
0.6
144
.049
+0.0
05
−0.0
0243
.833
+0.0
08
−0.0
082.46
+0.0
2−
0.0
223
.12+
6.4
0−
4.5
6a,
6
44.090
+0.0
02
−0.0
0543
.845
+0.0
08
−0.0
072.50
+0.0
2−
0.0
218
.66+
3.6
0−
2.7
76
44.068
+0.0
02
−0.0
0443
.851
+0.0
07
−0.0
062.45
+0.0
2−
0.0
215
.88+
2.4
2−
2.0
76
43.740
+0.0
04
−0.0
0443
.672
+0.0
06
−0.0
062.26
+0.0
1−
0.0
113
.01+
0.5
7−
0.5
06
43.623
+0.0
04
−0.0
06
43.65+
0.0
1−
0.0
12.11
+0.0
2−
0.0
211
.48+
1.6
9−
1.5
66
PG
1613
+65
8Mrk
876
0.12
945
.43
+0.00
Quiet−1.21
12.30
0.82
0+0.1
06
−0.0
92
12.25
8.34
+0.4
6−
0.5
1−0.64
+0.5
1−
0.4
644
.28+
0.0
6−
0.0
344
.38+
0.0
7−
0.0
91.95
+0.1
0−
0.1
028
.24+
107.7
8−
20.6
7a,
6
44.44+
0.0
3−
0.0
244
.43+
0.0
5−
0.0
62.12
+0.0
8−
0.0
810
.45+
10.7
8−
4.7
76
PG
1617
+17
5Mrk
877
0.11
444
.93
−0.14
Quiet−1.64
11.75
0.90
3+0.0
97
−0.0
81
11.55
8.67
+0.4
4−
0.4
5−1.48
+0.4
5−
0.4
4a
PG
1626
+55
4...
0.13
345
.17
−0.96
Quiet−1.37
11.84
0.97
6+0.0
24
−0.0
19
10.90
8.39
+0.6
1−
0.6
2−1.08
+0.6
2−
0.6
144
.17+
0.0
8−
0.0
844
.21+
0.0
7−
0.0
72.04
+0.1
5−
0.1
4<0.01
a,5,
6
PG
2130
+09
9II
Zw
136
0.06
344
.46
−0.49
Quiet−1.47
11.78
0.99
5+0.0
05
−0.0
10
11.63
7.43
+0.4
3−
0.4
3−0.17
+0.4
3−
0.4
343
.708
+0.0
11
−0.0
09
43.62+
0.0
2−
0.0
22.29
+0.0
5−
0.0
55.91
+0.7
3−
0.6
2j,6
PG
2214
+13
9Mrk
304
0.06
5844
.45
−1.30
Quiet−2.02
11.78
0.99
8+0.0
02
−0.0
04
11.46
8.44
+0.6
2−
0.6
2−1.18
+0.6
2−
0.6
243
.46+
0.0
4−
0.0
743
.64+
0.0
5−
0.1
71.80
+0.1
6−
0.1
64.48
+0.6
8−
0.6
8e,
6
PG
2233
+13
4...
0.32
6545
.94
−0.55
Quiet−1.66
12.56
1a12
.33
7.93
+0.6
1−
0.6
20.12
+0.6
2−
0.6
144
.52
2.41
+0.1
8−
0.1
8<0.01
e,2
PG
2349−01
44C−01
.61
0.17
4245
.51
12.59
0.90
4+0.0
45
−0.0
54
11.90
9.14
+0.5
0−
0.5
0−1.11
+0.5
0−
0.5
044
.57
1.78
+0.2
0−
0.3
5<0.01
e,5
Tab
le1continued
Hig
hly
Ionized
NV
and
OV
IO
utflo
ws
inth
eQ
UEST
Quasa
rs
7Table 1 (continued)
Name Other z log νLν(UV) log R Radio αOX log(LbolL
) αAGN log(LIRL
) log(MBHM
) log ηEdd log(LSX) log(LHX) ΓX NH Ref.
Name [erg s−1] Class [erg s−1] [erg s−1] [1022 cm−2](1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17)
Note—Column (1): Object name. Column (2): Other name. Column (3): Redshifts, with reference listed in Column (17). Where available,redshifts are based on the [O III] narrow line. For 3 quasars, we use H I (de Vaucouleurs et al. 1991; Carilli et al. 1998; Springob et al. 2005)instead; for two, full-spectrum fits to SDSS spectra (Schneider et al. 2010); and for a single case, CO data (Evans et al. 2006). Column (4):Logarithm of the monochromatic luminosity at rest-frame 1125 Å derived using the Galactic extinctions from Schlafly & Finkbeiner (2011) andthe reddening curve with RV = 3.1 of Fitzpatrick (1999). Column (5): logarithm of R, the ratio of radio-to-optical luminosity from Boroson& Green (1992). Column (6): Radio class − Quiet, Steep, or Flat depending on log R and radio spectral index from Boroson & Green (1992).Column (7): X-ray to optical spectral index αOX = 0.372 log (f2 keV/f3000 A) from Brandt et al. (2000), where f2 keV and f3000 A are therest-frame flux densities at 2 keV and 3000 Å, respectively. (For Mrk 231, we report the value from Teng et al. (2014).) Column (8): Bolometricluminosity in solar units calculated from 7 × L(5100 Å) + LIR (Netzer et al. 2007), where L(5100 Å) is the continuum luminosity λLλ at 5100 Årest wavelength and LIR is the 1 − 1000 µm infrared luminosity listed in column (10). We adopt a cosmology of H0 = 69.3 km s−1; Ωm = 0.287;Ωλ = 0.713 (WMAP9). Column (9): Fraction of the bolometric luminosity produced by the AGN, i.e. αAGN = LAGN/LBOL, based on the Spitzerresults (Veilleux et al. 2009a). The error bars are computed from the lowest and highest values among the six methods from this paper. Column(10): Logarithm of the 1 − 1000 µm infrared luminosity in solar units from Zhuang et al. (2018), except Mrk 231 (U et al. 2013), PG 1626+554(Lyu et al. 2017), and PG 2349−014 (Veilleux et al. 2009b). Column (11): Logarithm of the black hole mass in solar units from reverberationmapping (RM) measurements from The AGN Black Hole Mass Database (Bentz & Katz 2015) or, if unavailable, single-epoch measurements fromVestergaard & Peterson (2006), normalized down from f = 5.5 (Onken et al. 2004) to f = 4.3 to match RM scaling. According to Vestergaard &Peterson (2006), single-epoch measurements should have an extra 0.43 error added in quadrature. For PG 2349−014, we used the more uncertainphotometric measurement of Veilleux et al. (2009b). For 3C 273, we recorded the GRAVITY measurement (Gravity Collaboration et al. 2018).Column (12): Logarithm of the ratio of the bolometric luminosity to the Eddington luminosity. Column (13): Luminosity in the soft X-rays (0.5 −2 keV). Column (14): Luminosity of the hard X-rays (2 − 10 keV). Column (15): Photon index of the best-fit absorbed power-law distribution tothe X-ray emission (dN/dE ∝ E−ΓX where E is the X-ray photon energy). Column (16): Column density in units of 1022 cm−2. For most datafrom Teng & Veilleux (2010), this is from the best-fit absorbed power-law. Where this best fit is unabsorbed and Swift/BAT data are available,we substitute NH from Ricci et al. (2017). For X-ray related quantities, different rows = different observations, dates. Column (17): Referencesfor the redshift and X-ray measurements.
References—Redshift: (a) Boroson & Green 1992; (b) Carilli et al. 1998; (c) de Vaucouleurs et al. 1991; (d) Evans et al. 2006; (e) Hewett & Wild2010; (f) Ho & Kim 2009; (g) Hu et al. 2020; (h) Marziani et al. 1996; (i) Schneider et al. 2010; (j) Springob et al. 2005; X-ray: (1) Inoue et al.2007; (2) Jin et al. 2012; (3) Laha et al. 2018; (4) Piconcelli et al. 2005; (5) Ricci et al. 2017; (6) Teng & Veilleux 2010; (7) Teng et al. 2014; (8)Veilleux et al. 2014; (9) Waddell & Gallo 2020
aNo measurement; we assume a value of 1.
b Best-fit value is 0; we assume an upper limit of 1021 cm−2.
8 Veilleux et al.
0.0 0.1 0.2 0.3
11.5
12.0
12.5
13.0
0.0 0.1 0.2 0.3Redshift
11.5
12.0
12.5
13.0
log(
L bol/L
O •)
Figure 1. Bolometric (AGN + starburst) luminosities of theQUEST quasars in the sample as a function of their redshifts.
4. HST DATA
We obtained high-quality spectra for 19 quasarsusing the Cosmic Origins Spectrograph (COS) withgrating G130M under HST PID 12569 in Cycle 19(PI Veilleux). We acquired multi-epoch COS/G130Mdata on PG 1411+442 under programs 13451, 14460,and 14885 (PI Hamann). We searched for archivalCOS/G130M spectra of other quasars in the QUESTsample, as well as for multi-epoch data on the subsamplewe observed. We found archival data on 14 additionalQUEST quasars and recent multi-epoch exposures forthree quasars in the subsample first observed in Cycle19. We list the characteristics of these observations inTable 2.Since most of the Cycle 19 COS data have not yet
been the subject of a paper (the exceptions are Mrk 231and PG 1411+442; Veilleux et al. 2013b, 2016; Hamannet al. 2019b), we briefly summarize here how they wereobtained. A total of 24 orbits were allocated for these19 targets with most targets requiring one orbit. Theexceptions are PG 1004+130 (2 orbits) and Mrk 231(5 orbits). All but Mrk 231 are point sources withaccurate positions; they were acquired directly usingACQ/PEAKXD and ACQ/PEAKD. For Mrk 231, aNUV image was obtained with ACQ/IMAGE. All obser-vations were observed in time-tag mode to allow us to ex-clude poor quality data and improve thermal correctionand background removal. We split the exposures intofour segments of similar durations at two FP_POS set-tings (#2 and #4) and two wavelength settings (CEN-
WAVE) separated by ∼20 Å. This observing strategyreduces the fixed pattern noise and fills up the chip gapwithout excessive overheads.The observations include at least 1150–1450 Å in the
observer’s frame. This range includes redshifted O VIλλ 1032, 1038, N V λλ1238, 1243, Lyα λ1216 and/orLyβ λ1025 in emission and/or absorption. In at leasttwo cases (PG 1126−041, PG 1411+442, and perhapsalso PG 1001+054 and PG 1004+130; see Sec. 7.4), theweaker P V λλ1117, 1128 absorption lines are also de-tected. The specific lines covered depend on the quasarredshift. The short-wavelength cutoff of the COS pre-vents us from searching for O VI systems in quasars withz . 0.11, while N V systems are redshifted out of theCOS data in quasars with z & 0.18. It is therefore pos-sible to study both O VI and N V only over a limitedrange of quasar redshifts. Nevertheless, we achieve ourscience goals by covering at least one H I Lyman seriesline and one high-ionization doublet (O VI and/or N V).In at least two cases (PG 1411+442 and PG 1004+130),weaker and/or lower-ionization lines, such as C II λ1335,C III λ977, N III λ990, O I λ1304, Si II λ1260, Si IIIλ1206, and Si IV λλ1394, 1403, are also present in thespectra. These lines may be used to help constrain thelocation, ionization, total column densities (NH) andmetal abundances in the absorbing gas (e.g. Hamannet al. 2019b). PG 1004+130, one of the highest-redshiftsources in our sample, also shows S VI λλ933, 945.All of our sample have data with CENWAVE of 1291,
1300, 1309, 1318, and/or 1327, and almost all of thedata (with the exception of the final round of data onPG 1001+054) were obtained in COS lifetime positions(LP) 1–3. For these CENWAVE settings, the spectralresolution of COS increases with wavelength and hasdegraded somewhat with changes in LP, but is still >104
at all wavelengths. This corresponds to resolution betterthan 30 km s−1 FWHM at all wavelengths, ranging up toa peak of ∼15 km s−1 at LP1 and 1450 Å. Three quasarshave additional data from CENWAVE 1055, 1096, or1222 observations. For these CENWAVE values, thespectral resolution peaks in the FUVB (blue) segmentat >104, but is lower in FUVA, with average values of∼3000, 5000, and 104, respectively.We downloaded all exposures from the Hubble Legacy
Archive and determined that they were processed byCALCOS v3.3.10. For each quasar, we coadded all ex-posures with CENWAVE 1222–1327 into a single spec-trum using v3.3 of coadd_x1d.pro (Danforth et al. 2010),setting BIN=3. The resulting median S/N per binnedpixel over 1290–1310 Å is 11.5, with a standard devia-tion of 10.9 and a range of 2–52. (The low end of therange arises in a strong N V BAL in PG 1126−041.) We
Highly Ionized N V and O VI Outflows in the QUEST Quasars 9
Table 2. Summary of the Observations
Name Range [Å] CENWAVE [Å] texp[s] Date PID PI
(1) (2) (3) (4) (5) (6) (7)
PG 0007+106 1151–1470 1309/1327 1868 2011-12-14 12569 S. VeilleuxPG 0026+129 1133–1451 1291/1309 1868 2011-10-25 12569 S. VeilleuxPG 0050+124 1151–1470 1309/1327 1868 2012-11-01 12569 S. Veilleux
1133–1465 1291/1309/1327 7621 2015-01-20 13811 E. CostantiniPG 0157+001 1151–1469 1309/1327 1828 2012-01-25 12569 S. VeilleuxPG 0804+761 1136–1458 1291/1300/1309/1318 5510 2010-06-12 11686 N. AravPG 0838+770 1136–1458 1291/1300/1309/1318 8865 2009-09-24 11520 J. GreenPG 0844+349 1151–1470 1309/1327 1900 2012-03-06 12569 S. VeilleuxPG 0923+201 1133–1451 1291/1309 1860 2012-03-14 12569 S. VeilleuxPG 0953+414 1136–1458 1291/1300/1309/1318 4785 2011-10-18 12038 J. GreenPG 1001+054 1066–1367 1222 2068 2014-04-04 13423 R. Cooke
1140–1455 1291/1300/1309/1318 3165 2014-06-19 13347 J. Bregman1131–1429 1291 2902 2019-03-26 15227 J. Burchett
PG 1004+130 1133–1451 1291/1309 4107 2011-12-21 12569 S. VeilleuxPG 1116+215 1136–1458 1291/1300/1309/1318 4677 2011-10-25 12038 J. GreenPG 1126−041 1152–1470 1309/1327 1856 2012-04-15 12569 S. Veilleux
901–1200 1055 1874 2014-06-01 13429 M. Giustini1171–1467 1327 1580 2014-06-01 13429 M. Giustini900–1200 1055 1874 2014-06-12 13429 M. Giustini1171–1467 1327 1580 2014-06-12 13429 M. Giustini900–1200 1055 1874 2014-06-28 13429 M. Giustini1171–1467 1327 1580 2014-06-28 13429 M. Giustini901–1200 1055 1837 2015-06-14 13836 M. Giustini1171–1468 1327 1540 2015-06-14 13429 M. Giustini
PG 1211+143 1171–1472 1327 2320 2015-04-14 13947 J. LeePG 1226+023 1135–1470 1291/1300/1309/1318/1327 4002 2012-04-22 12038 J. GreenPG 1229+204 1152–1469 1309/1327 1868 2012-04-26 12569 S. Veilleux
Mrk 231 1152–1472 1309/1327 12536 2011-10-15 12569 S. VeilleuxPG 1302−102 1136–1458 1291/1300/1309/1318 5979 2011-08-16 12038 J. GreenPG 1307+085 1152–1470 1309/1327 1836 2012-06-16 12569 S. VeilleuxPG 1309+355 1133–1451 1291/1309 1896 2011-12-06 12569 S. VeilleuxPG 1351+640 1152–1470 1309/1327 2108 2011-10-21 12569 S. VeilleuxPG 1411+442 1152–1470 1309 1936 2011-10-23 12569 S. Veilleux
941–1241 1096 4954 2015-02-12 13451 F. Hamann1152–1453 1309 1917 2015-02-12 13451 F. Hamann941–1241 1096 2407 2016-04-16 14460 F. Hamann1152–1453 1309 1954 2016-04-16 14460 F. Hamann941–1241 1096 1783 2017-06-10 14885 F. Hamann1152–1453 1309 1847 2017-06-10 14885 F. Hamann
PG 1435−067 1133–1451 1291/1309 1864 2012-02-29 12569 S. VeilleuxPG 1440+356 1152–1470 1309/1327 1924 2012-01-26 12569 S. VeilleuxPG 1448+273 1136–1448 1291/1309 2946 2011-06-18 12248 J. TumlinsonPG 1501+106 1132–1434 1291 3121 2014-07-06 13448 A. FoxPG 1613+658 1145–1467 1300/1309/1318/1327 9499 2010-04-08 11524 J. Green
1133–1429 1291 3080 2010-04-09 11686 N. AravPG 1617+175 1133–1451 1291/1309 1844 2012-06-16 12569 S. VeilleuxPG 1626+554 1136–1458 1291/1300/1309/1318 3318 2011-06-15 12029 J. GreenPG 2130+099 1135–1458 1291/1300/1309/1318 5513 2010-10-28 11524 J. GreenPG 2214+139 1152–1463 1309/1327 1401 2011-11-08 12569 S. Veilleux
1138–1434 1291 2082 2012-09-21 12604 A. FoxPG 2233+134 1171–1472 1327 2104 2014-06-18 13423 R. CookePG 2349−014 1152–1470 1309/1327 1844 2011-10-20 12569 S. Veilleux
Note—Column (1): Name of object; Column (2): Wavelength range, in Å; Column (3): CENWAVE setting(s);Column (4): Exposure time, in seconds; Column (5): Start date; Column (6): Proposal ID; Column (7):Program principal investigator.
separately coadded the two quasar datasets with CEN-WAVE 1055 and 1096.
5. DATA ANALYSIS
We conducted a uniform analysis of the high-ionization absorbers in our sample. In this section, wedescribe the methods we used to identify and character-ize these absorption features.We used v0.5 of the publicly available, IDL-based IFS-
FIT package (Rupke 2014; Rupke & Veilleux 2015) to
model the absorption lines. The rest of the softwareused to model the data (continuum fitting, plotting,regressions) is contained in or called from our publicCOSQUEST repository on GitHub (Rupke 2021a).
5.1. Model Fitting
The starting point of the analysis is to identify thevarious emission and absorption lines produced by thequasars and their environments. Since our program isfocused on QSO and ULIRG outflows, we only identify
10 Veilleux et al.
and measure absorption lines within ∼10,000 km s−1 ofthe QSO redshifts. We refer to these lines as “associ-ated” absorbers. Identifications of the foreground “inter-vening” absorbers can be found in Tripp et al. (2008),Savage et al. (2014), and Danforth et al. (2016). We firstcompare each quasar spectrum against a list of commonUV absorbers in quasar spectra (Prochaska et al. 2001,Figure 2). We list the quasar redshifts in Table 1. Mostof these redshifts are derived from narrow optical emis-sion lines ([O III], Hβ) and may underestimate the truerecession velocities since some fraction of the line emis-sion may arise from the outflowing material itself (e.g.Rupke et al. 2017). See Teng et al. (2013) for a compar-ison of these measurements with the H I 21-cm emissionand absorption line profiles.
PG1126−041, z = 0.060
1200 1250 1300 1350 1400 1450Observed Wavelength (Å)
0.0
5.0
10.0
15.0
Fλ/1
0−14 (
ergs
s−
1 cm
−2 Å
−1 )
1160 1170 1180 1190 12000.00.51.01.52.02.53.03.5
PII
1152
CI 1
157
FeI
I 109
6
PV
111
7
FeI
I 112
1F
eIII
1122
SIII
119
0, S
iII 1
190
FeI
I 112
5S
iII 1
193
PV
112
8
NI 1
199
NI 1
200
NI 1
200
NI 1
134,
NI 1
134
NI 1
134
1210 1220 1230 1240 12500.0
1.0
2.0
3.0
SiII
I 120
6
FeI
I 114
3
FeI
I 114
5
Lyα
PII
1152
CI 1
157
NV
123
8
CI 1
139
NV
124
2
SII
1250
SII
1253
1260 1270 1280 1290 1300 13100.02.04.06.08.0
10.012.014.0
SII
1259
SiII
126
0, F
eII 1
260
SIII
119
0, S
iII 1
190
SiII
119
3
NI 1
199
NI 1
200
NI 1
200
CI 1
277
SiII
I 120
6
CI 1
280
Lyα
PII
1301
, OI 1
302
SiII
130
4
1320 1330 1340 1350 13600.01.02.03.04.05.0
NV
123
8
CI 1
139
NiII
131
7, N
V 1
242
SII
1250
CI 1
328,
SII
1253
CII
1334
SII
1259
CII*
133
5S
iII 1
260,
FeI
I 126
0
CI 1
277
OI 1
355
CI 1
280
1370 1380 1390 1400 14100.01.0
2.0
3.0
4.0
NiII
137
0
PII
1301
OI 1
302
SiII
130
4
SiIV
139
3
NiII
131
7
SiIV
140
2
CI 1
328
CII
1334
CII*
133
5
1420 1430 1440 1450 14600.01.0
2.0
3.0
4.0
OI 1
355
NiII
137
0
NiII
145
4
NiII
146
7N
iII 1
467
Figure 2. An example of a FUV spectrum used in the study.Shown here is the COS spectrum of PG 1126−041, wherethe data are displayed in black and the expected positionsof the features in the Milky Way and quasar rest-frames areindicated in blue and red, respectively.
Next, we fit the continuum and broad line emission(Lyα, N V, and O VI) in three separate spectral win-dows around Lyα, N V, and O VI+Lyβ. In the twoquasars in which we fit P V, this continuum region isalso fit separately. Within each of these windows, weuse a piecewise function of 1–4 segments in the major-ity of cases. In relatively featureless spectral regions,these segments are low-order polynomials. In more com-
plex spectral regions we employ cubic B-splines. TheB-splines are themselves piecewise polynomials, and weseparate the spline knots by a typical interval of 3 Å. Weinvoke BSPLINE_ITERFIT from the SDSS IDLUTILSlibrary to fit the B-splines.In seven cases, the fits with piecewise functions are
poorly constrained. For PG 1001+054, PG 1004+130,PG 1411+442, and PG 1617+175, PG2130+099, andPG2214+139 this is due to broad, deep absorption fea-tures over which it is difficult to fit polynomials orsplines. For a seventh quasar—PG 1351+640—the poorconstraints are due to several narrow absorbers near thepeak of Lyα. In two of these cases, we instead use aLorentzian profile to fit Lyα. For the five others we usethe BOSS template from Harris et al. (2016), scale itmultiplicatively by a low-order power law, and add alinear pedestal.After fitting the continuum, we normalize the data in
each spectral window by dividing by this fit.We characterize the doublet absorption features (N V
λλ1238, 1243; O VI λλ1032, 1038; and P V λλ1117,1128) in the quasar spectra using simple model fits. Ourprimary objectives are to estimate the overall equivalentwidths and kinematics of the outflowing gas associatedwith these features (mass, momentum, and energy esti-mates are beyond the scope of the present paper, exceptfor a few special cases discussed in Section 7.4). Weare not aiming to derive precise column densities fromthe (often saturated) absorption line profiles, so the useof the precise COS line-spread function (LSF) is not re-quired here (we return to this point at the end of this sec-tion). If the lines within these doublets were unblended,fits to the intensity profiles of the individual lines wouldthus be sufficient. However, the doublet lines are oftenstrongly blended because of (1) strong blueshifts due tohigh outflow velocities and (2) broad line profiles dueto multiple clouds along the line of sight and/or largelinewidths. We thus adopt the doublet fitting procedureof Rupke et al. (2005), which is optimized for blendeddoublets. In this method, the total absorption profilesof a feature are fit as the product of multiple doubletcomponents. Each component is a Gaussian in opticaldepth τ vs wavelength with a constant covering factorCf . Within each doublet the two lines have a constantτ ratio. This allows us to simultaneously fit τ and Cf ,which are otherwise degenerate in the fit of a single line.The free parameters in the fit to each doublet compo-nent are thus Cf , peak τ , velocity width, and centralwavelength. The determination of the number of com-ponents needed in the fit is subjective and non-linear− it depends on the line complexity and data quality.The main goal here is to get a good fit to the absorption
Highly Ionized N V and O VI Outflows in the QUEST Quasars 11
PG1126−041, z=0.060
−5000 −4000 −3000 −2000 −1000 0 1000Velocity (km/s)
0.00.2
0.4
0.6
0.8
1.0
Nor
mal
ized
Fλ
PV 1117
0.00.2
0.4
0.6
0.8
1.0
PV 1128
0.00.2
0.4
0.6
0.8
1.0
OVI 1032
0.00.2
0.4
0.6
0.8
1.0
OVI 1038
0.00.2
0.4
0.6
0.8
1.0
NV 1238
0.00.2
0.4
0.6
0.8
1.0
NV 1242
0.00.2
0.4
0.6
0.8
1.0
Lyα
0.00.2
0.4
0.6
0.8
1.0
Lyβ
Figure 3. An example of interline comparison that is usedto identify absorbing systems associated with the quasars.The results shown here are for PG 1126−041, produced bydividing the spectrum shown in Fig. 2 by a smooth poly-nomical/spline/template fit to the continuum near the keyabsorption lines of our study and plotting the results in ve-locity space in the quasar rest frame. The data are in black,the components used to fit the absorption profiles are shownin blue, and the overall fit is shown in purple. The veloc-ity centroids of the main absorbing systems are indicated byblue (O VI) or red (Lyα, Lyβ, N V, P V) vertical dottedlines. The two strong lines in the panel labeled P V λ1128are Galactic ISM features without counterparts in P V λ1117or any of the other lines.
features to derive the equivalent widths and kinemat-
ics of the outflowing gas associated with these features.We do not attach a physical meaning to the individualcomponents in the fit.The general expression for the normalized intensity of
a doublet component is
I(λ) = 1− Cf + Cfe−τlow(λ)−τhigh(λ), (1)
where Cf is the line-of-sight covering factor (or the frac-tion of the background source producing the continuumthat is covered by the absorbing gas; though scatter-ing into the line of sight can also play a role) and τlow
and τhigh are the intrinsic optical depths of the lower-and higher-wavelength lines in the doublet (Rupke et al.2005). The background light source is assumed to bespatially uniform. The covering factor is the same forboth lines of the doublet. The peak (and total) opticaldepths of the resonant doublet lines in O VI, N V, andP V are related by a constant factor τlow/τhigh = 2.00
because of the 4-fold degeneracy in the upper state ofthe higher energy transition compared to the 2-fold de-generacy in the lower state. (The higher degeneracy isdue in turn to its higher total angular momentum quan-tum number j). For more than one doublet component,we use the product of the intensities of the individualcomponents, which is the partially-overlapping case ofRupke et al. (2005).Because the doublet profile shape–i.e., relative depths
of the two lines and trough shape–does not change sig-nificantly above optical depths τhigh of a few, we set alimit of τhigh ≤ 5. Out of 59 O VI components, 19 haveτhigh = 5, or 32%. For N V, 13 of 62 components haveτhigh = 5, or 21%.The results from these fits are also used to calculate
the total velocity-integrated equivalent widths of the ab-sorbers in the object’s rest frame,
Weq =
∫[1− f(v)]dv, (2)
the weighted average outflow velocity,
vwtavg =
∫v[1− f(v)]dv
Weq, (3)
and the weighted outflow velocity dispersion,
σrms =
(∫(v − vwtavg)2[1− f(v)]dv
Weq
) 12
, (4)
a measure of the second moment in velocity space of theabsorbers in each quasar. These quantities are similarto those defined by Trump et al. (2006), but withoutthe constraints on depth, width, or velocity. These con-straints have little effect on the results for our sample,
12 Veilleux et al.
but we find it useful to include possibly inflowing ab-sorbers. Note that Weq, vwtavg, and σrms are not cor-rected for partial covering. To test the impact of thisassumption on our results, we have recomputed themafter changing the absorption lines so that they haveCf = 1 instead of the measured Cf , and then redid theregression analysis discussed in Sec. 5.2. Only very smallchanges of order 1% in the p-values are observed if wecorrect for partial covering.Figure 3 shows the fits to the spectrum presented in
Figure 2. The fits to all of the features detected in theFUV spectra of the 33 quasars in our sample are pre-sented in Appendix A, and the results derived from thesefits are tabulated in Table 3.We computed errors in best-fit parameters and derived
model quantities by refitting the model spectrum 1000times. In each case we added Gaussian-distributed ran-dom errors to each pixel in the model with σ equal to themeasurement error. These formal errors are small dueto the high S/N in our data. Errors due to continuumplacement are likely to dominate the true error budget.We also estimated upper limits to the doublet equiva-
lent width in cases where we did not detect N V and/orO VI. To do so, we assumed an optically-thick (τ1243
or τ1038 = 5), v = 0, σ = 50 km s−1 absorption line.We set the covering factor equal to half the root-mean-square deviation in the continuum within ±0.5 Å of theexpected rest-frame location of each line in the doublet.(The factor-of-2 accounts for fitting 2 lines instead of 1.)We set the limit equal to the resulting model equivalentwidth.The optical depths and covering factors derived from
our fitting scheme are approximations. Though itis a physically-motivated way to decompose strongly-blended doublets, the method implicitly assumes thatthe velocity dependences of Cf and τ can be describedas the sum of discrete independent Gaussians. In re-ality, they are probably more complex functions of ve-locity (e.g. Arav et al. 2005, 2008). In several cases–the N V absorbers in PG 1001+054, PG 1411+442,PG 1617+175, and PG2214+139, and the O VI ab-sorbers in PG 1001+054 and PG 1004+130–the fitsinclude very broad components that cannot be distin-guished from complexes of narrower lines given the dataquality. In two O VI absorbers (PG 0923+201 andPG 1309+355), there are no data on the blue line be-cause it is contaminated by geocoronal Lyα, so any con-straints on τ and Cf come solely from line shape. Fi-nally, in four O VI fits (PG 1001+054, PG 1004+130,PG 1126−041, and PG 1617+175), the Lyβ and O VIabsorption lines blend together and cannot be easilyseparated in the fit. In three of these cases (all but
PG 1001+054), we simply fit the visible absorption asdue solely to O VI at wavelengths in which there is atleast some O VI absorption contributing to the spec-trum. For the fourth case, we are able to roughly sep-arate the lines by fitting only down to a specific wave-length. A detailed object-by-object discussion is givenin Appendix A.Despite these caveats, the fitting procedure is suffi-
cient to meet our primary objectives of estimating theoverall equivalent widths and kinematics of these fea-tures. The 3-σ detection limit on the doublet equivalentwidths is typically ∼ 20 mÅ in our data although itvaries from one spectrum to the other.We have conducted detailed tests of the impact of the
COS LSF on our measurements to verify that the useof the precise COS LSF is not required here. In oneseries of simulations, we created a series of fake, sat-urated Voigt line profiles with a median S/N of 5 perpixel and line widths ranging from σ = 10 km s−1 to 50km s−1. We convolved the profiles with the LSF down-loaded from the COS website. We find that the LSFcauses a difference of up to only ∼10% on the line widthand covering fraction measurements for lines with σ ≥20 km s−1 (the corresponding Doppler b parameter ofthe Voigt profile is
√2 σ ' 28 km s−1), which is smaller
than the values measured for nearly all of the absorbersdetected in our objects (Table 3). We have also run aCOS LSF analysis on a ULIRG with narrow N V absorp-tion features, taken from the sample of Paper II. Usingthe method described here, we get a Doppler b param-eter of 78 km s−1 and covering fraction of 0.84, whilethe COS LSF gives 83 km s−1 and 0.82, respectively,confirming that the results for the relatively broad ab-sorbers reported in the present paper are reliable.
5.2. Regressions
To search for connections between outflow andquasar/host properties, we computed linear regressionsbetween the properties in Tables 1 and 3. In mostcases, we apply the Bayesian model in LINMIX_ERR(Kelly 2007). We use the Metropolis-Hastings samplerand a single Gaussian to represent the distribution ofquasar/host parameters (except for Weq vs. AGN frac-tion, for which we used NGAUSS=3). LINMIX_ERRpermits censored y-values, which is the case for Weq.When we compute the regressions for the indepen-
dent variable NH, however, the x-axis values are alsocensored. In this case we turn to the method of Isobeet al. (1986) for computing the Kendall tau correlationcoefficient with censored data in both axes. We use theimplementation of pymccorrelation (Privon et al. 2020),which in turn perturbs the data in Monte Carlo fash-
Highly Ionized N V and O VI Outflows in the QUEST Quasars 13
Table 3. Results from the Multi-Component Fits to the Absorbers
Name Line Weq vwtavg σrms # comp.
Å km s−1 km s−1
(1) (2) (3) (4) (5) (6)
PG0007+106 N V <0.16PG0026+129 N V <0.09
O VI <0.12PG0050+124 N V 0.88+0.019
−0.017 -1106.3+19.0−18.7 595.0+18.0
−16.4 6PG0157+001 N V <0.13
O VI <0.09PG0804+761 N V 0.02+0.003
−0.003 591.0+1.8−1.7 12.3+1.9
−1.8 1O VI 0.18+0.009
−0.007 571.0+1.6−1.5 27.8+1.1
−1.1 1PG0838+770 N V <0.12
O VI <0.06PG0844+349 N V 0.73+0.011
−0.011 151.4+0.8−0.8 31.8+0.5
−0.5 2PG0923+201 O VI 4.48+0.072
−0.072 -3048.3+11.4−11.9 335.6+9.3
−8.7 1PG0953+414 O VI 0.20+0.006
−0.006 -825.4+14.7−13.9 418.3+7.2
−7.1 2PG1001+054 N V 6.07+0.058
−0.059 -5969.9+24.0−37.4 1326.5+18.0
−11.7 4O VI 11.61+0.080
−0.072 -5743.9+33.2−83.1 1092.4+32.7
−27.6 6PG1004+130 O VI 24.29+0.245
−0.277 -5335.7+80.0−75.4 2970.2+72.3
−47.4 12PG1116+215 O VI 0.27+0.008
−0.008 -2271.6+43.2−46.3 908.7+22.1
−25.6 2PG1126-041 N V 10.66+0.037
−0.034 -2085.5+10.0−9.8 1076.1+6.9
−7.2 13O VI 16.80+1.891
−1.616 -2559.2+298.8−251.5 1482.2+132.6
−169.4 10P V 0.21+0.017
−0.017 -2234.2+5.9−6.1 48.2+4.3
−3.8 2PG1211+143 N V <0.05PG1226+023 N V <0.04
O VI <0.04PG1229+204 N V <0.09Mrk 231 N V <0.18PG1302-102 O VI <0.05PG1307+085 N V <0.10
O VI 0.15+0.017−0.016 -3406.2+11.9
−14.1 66.8+11.8−14.1 2
PG1309+355 O VI 8.57+0.038−0.031 -893.9+4.9
−4.7 364.5+3.2−3.1 7
PG1351+640 N V 4.36+0.036−0.034 -1264.4+10.0
−11.4 428.4+4.8−3.8 9
PG1411+442 N V 10.30+0.019−0.018 -1594.8+2.6
−2.4 562.7+2.5−2.3 4
P V 0.83+0.028−0.028 -1754.6+5.9
−5.5 131.0+3.8−3.6 2
PG1435-067 N V <0.15O VI <0.08
PG1440+356 N V 0.89+0.023−0.021 -1478.4+28.1
−26.7 775.3+27.0−25.8 3
PG1448+273 N V 3.22+0.038−0.033 -229.8+2.2
−2.1 164.1+1.0−1.0 4
PG1613+658 N V 0.14+0.005−0.004 -3714.6+6.1
−5.5 121.8+3.2−3.6 2
O VI 0.68+0.008−0.008 -3691.1+2.0
−1.8 126.6+0.8−0.8 2
PG1617+175 N V 3.00+0.059−0.055 -3094.7+19.9
−23.8 526.8+34.3−22.5 5
O VI 6.08+0.173−0.173 -3323.7+113.3
−133.0 920.2+83.2−71.6 8
P V 0.06+0.040−0.033 -3355.0+55.1
−38.0 42.1+46.1−25.8 1
PG1626+554 N V <0.08O VI <0.06
PG2130+099 N V 0.76+0.013−0.013 -1312.3+9.9
−9.6 540.3+9.5−9.3 3
PG2214+139 N V 8.01+0.024−0.023 -1461.1+17.9
−17.9 681.7+21.9−20.2 5
PG2233+134 O VI 0.17+0.025−0.024 -211.2+3.0
−3.0 17.3+3.2−2.9 1
PG2349-014 N V <0.12
Note—Column (1): Name of object. Column (2): N V means N V λ1238, 1243,O VI means O VI λ1032, 1038, and P V means P V λ1117, 1128. N V orO VI is not listed when it lies outside of the spectral range of the data. Column(3): Velocity-integrated equivalent widths (eq. 2). Column (4): Average depth-weighted outflow velocity (eq. 3), which is a measure of the average velocity ofthe outflow systems in each object. Column (5): Average depth-weighted outflowvelocity dispersion (eq. 4), which is a measure of the range in velocity of theoutflow systems in each quasar. Column (6): Number of absorption components.
14 Veilleux et al.
ion to compute the errors in the correlation coefficient(Curran 2014).For both regression methods, we computed the signifi-
cance of a correlation as the fraction of cross-correlationvalues r < 0 (r > 0) for a positive (negative) best-fit r.For LINMIX_ERR, the r values are draws from the pos-terior distribution, while for pymccorrelation they areresults of the Monte Carlo perturbations.We do not consider the N V and O VI points inde-
pendent for the purposes of the regressions. Therefore,where both doublets are present in the data for a givenquasar, we compute the average measurement (eitherdetection or limit) from the two lines. If only one lineis detected, we use that measurement rather than av-eraging a detection and a limit. Where multiple X-raymeasurements exist for a quasar, we take the average.Errors in LBOL, LIR/LBOL, LFIR/LBOL, and αOX areunknown, so for the purposes of regression we fix theerrors to 0.1 dex. For νLν(UV), we ignore the negligiblestatistical measurement errors.
6. RESULTS
The results from our spectral analysis of the HST spec-tra are summarized in Table 3. In this section, we inves-tigate whether the presence or nature of quasar-drivenoutflows and starburst winds correlate with the prop-erties of the quasars and host galaxies. The quantitiesthat we consider in our correlation matrix are listed inTable 1 and defined in the notes to that table. Theresults from the statistical and regression analyses aresummarized in Tables 4 and 5.Note that we do not make a distinction between
quasar-driven outflows and starburst-driven winds inthis section, and we only consider absorption lines within10,000 km s−1 of the QSO redshift (inclusion of lines atgreater displacements leads to unacceptable contamina-tion by intervening absorbers). A more comprehensiveassessment of the detected outflows is conducted in Sec-tion 7, after we have considered the line profiles morefully, including signs of saturation and/or partial cover-ing of the continuum source (Sec. 6.2) and the overallkinematics of the outflowing gas (Sec. 6.4). Similarly,the comparison of our results with those from previousstudies is postponed until Section 7, once the resultsfrom our spectral analysis have been fully presented.
6.1. Rate of Incidence of Outflows
Figures 4 and 5 show the median velocities in thequasar rest frame of all of the detected N V and O VIabsorption-line systems, sorted from top to bottom bydecreasing redshift and bolometric luminosity, respec-tively. The first of these figures clearly illustrates the
fact mentioned in Section 4 that our ability to detectthe N V and O VI features is limited by the spectralcoverage of the COS data to z . 0.18 and z & 0.11,respectively (systems outside of the spectral range areindicated by an “x” in this figure and Fig. 5).
0 −2000 −4000 −6000 −8000v50 (km/s)
PG1501+106
Mrk 231
PG0050+124
PG1126−041
PG2130+099
PG0844+349
PG1229+204
PG1448+273
PG2214+139
PG1440+356
PG1211+143
PG1351+640
PG0007+106
PG1411+442
PG0804+761
PG1617+175
PG1613+658
PG1435−067
PG0838+770
PG1626+554
PG0026+129
PG1307+085
PG1226+023
PG1001+054
PG0157+001
PG2349−014
PG1116+215
PG1309+355
PG0923+201
PG0953+414
PG1004+130
PG1302−102
PG2233+134
Qua
sars
, sor
ted
by z
NV OVI00
NV 1238,1243O VI 1032,1038NV, σ < 25 km/sO VI, σ < 25 km/s
not in spectral rangeaffected by geocoronal line or chip gapundetected
Figure 4. Median velocities of the N V and O VI absorbingsystems detected in the QUEST quasars of our sample. Notethat the faster outflows with more negative velocities lie onthe right in this figure. The objects are sorted from top tobottom in order of decreasing redshift. Red symbols markN V λλ1238, 1243 and blue symbols mark O VI λλ1032,1038. Open symbols indicate systems with velocity disper-sion σ < 25 km s−1. The two columns on the right indicatewhether N V (red) or O VI (blue) is within the spectralrange of the data (“x” indicates that it is not), affected bygeocoronal line or chip gap (encircled “x”), or simply unde-tected (downward-pointing triangle). A lack of symbol marksa detection.
A cursory examination of Figures 4 and 5 shows thatblueshifted N V or O VI absorption systems suggestiveof outflows (with equivalent widths above our 3-σ detec-tion limit of ∼ 20 mÅ) are detected in about 60% of thequasars in our sample, and there is no obvious trend inthe rate of incidence with redshift or bolometric lumi-nosity.The results of a more quantitative analysis based on β
distributions (Cameron 2011) are listed in Table 4. The
Highly Ionized N V and O VI Outflows in the QUEST Quasars 15
0 −2000 −4000 −6000 −8000v50 (km/s)
PG1501+106
PG1448+273
PG0844+349
PG1126−041
PG1229+204
PG1617+175
PG0838+770
PG2130+099
PG2214+139
PG1411+442
PG1440+356
PG1626+554
PG1001+054
PG1435−067
PG1211+143
PG1351+640
PG0050+124
PG0026+129
PG0804+761
PG0007+106
PG1613+658
PG1309+355
PG1307+085
PG0923+201
PG0953+414
PG1116+215
PG2233+134
PG2349−014
Mrk 231
PG1004+130
PG0157+001
PG1302−102
PG1226+023Q
uasa
rs, s
orte
d by
Lbo
l
NV OVI00
NV 1238,1243O VI 1032,1038NV, σ < 25 km/sO VI, σ < 25 km/s
not in spectral rangeaffected by geocoronal line or chip gapundetected
Figure 5. Same as Fig. 4, but the objects are sorted fromtop to bottom in order of decreasing bolometric luminosity.
overall rate of incidence of N V or O VI absorbers is61% with a 1-σ range of (52% − 68%), once taking intoaccount the spectral coverage of the data. This rate isvirtually the same for N V and O VI. Among quasarswith log LBOL/L > 12.0, this rate is 61% with a 1-σrange of (49% − 71%), while it is 60% (47% − 71%)among the systems of lower luminosities. These ratesare thus not significantly different from each other, andare similar to the rate of incidence of O VI outflows inlocal Seyfert 1 galaxies (Kriss 2004a,b) as well as C IV(Crenshaw et al. 1999) or X-ray (Reynolds 1997; Georgeet al. 1998) absorption.We have also searched for trends between the rate
of incidence of outflows and several other quantities.The detection rate of outflows (79%) among quasarsthat have strongly absorbed X-ray continua (NH > 1022
cm−2) is significantly higher than those that do not(25%) (Table 4). These rates differ at the 2-σ level(95.4%), where the ranges of the incidence rate are 56− 92% for quasars with absorbed X-ray continua and 9− 54% for the others. Using the scipy.stats imple-mentation of the Fisher exact test, the null hypothesisthat galaxies with strongly- and weakly-absorbed X-raycontinua UV absorbers are equally likely to show N Vor O VI absorbers is rejected at the 99.2% level. The
Table 4. Rate of Incidence of Outflows
Line Detection Total Fraction (1-σ range)
(1) (2) (3) (4)
All Quasars
N V 13 27 0.48 (0.39 − 0.58)O VI 12 20 0.60 (0.49 − 0.70)Both 5 14 0.36 (0.25 − 0.50)Any 20 33 0.61 (0.52 − 0.68)
log LBOL/L ≥ 12.0
N V 4 12 0.33 (0.23 − 0.48)O VI 9 14 0.64 (0.50 − 0.75)Both 2 8 0.25 (0.16 − 0.44)Any 11 18 0.61 (0.49 − 0.71)
log LBOL/L < 12.0
N V 9 15 0.60 (0.47 − 0.71)O VI 3 6 0.50 (0.32 − 0.68)Both 3 6 0.50 (0.32 − 0.68)Any 9 15 0.60 (0.47 − 0.71)
NH > 1022 cm−2
N V 10 16 0.62 (0.50 − 0.73)O VI 8 8 1.00 (0.81 − 0.98)Both 3 5 0.60 (0.38 − 0.76)Any 15 19 0.79 (0.67 − 0.85)
NH ≤ 1022 cm−2
N V 2 10 0.20 (0.13 − 0.37)O VI 2 10 0.20 (0.13 − 0.37)Both 1 8 0.12 (0.08 − 0.32)Any 3 12 0.25 (0.17 − 0.41)
αox ≥ −1.6
N V 7 17 0.41 (0.31 − 0.53)O VI 6 12 0.50 (0.37 − 0.63)Both 2 9 0.22 (0.14 − 0.41)Any 11 20 0.55 (0.44 − 0.65)
αox < −1.6
N V 6 9 0.67 (0.49 − 0.78)O VI 6 7 0.86 (0.64 − 0.91)Both 3 4 0.75 (0.48 − 0.85)Any 9 12 0.75 (0.59 − 0.83)
Note—Column (1): Feature(s) used in the statisticalanalysis. “Both” means both N V and O VI doubletsand “Any” means either N V or O VI doublet orboth; Column (2): Number of objects with detectedoutflows; Column (3): Number of objects in totalwith the appropriate redshift; Column (4): Fractionof objects with detected outflows. The two numbersin parentheses indicate the 1-σ range (68% probabil-ity) of the fraction of objects with detected outflows,computed from the β distribution (Cameron 2011).
16 Veilleux et al.
Table 5. Linear Regression Results
y x N p r
(1) (2) (3) (4) (5)
Weq log(LBOL/L) 32 0.046 -0.34+0.19−0.17
Weq log[λL1125/erg s−1] 32 0.140 -0.22+0.21−0.18
Weq AGN fraction 32 0.094 0.78+0.19−0.61
Weq log(LAGN/L) 32 0.054 -0.33+0.19−0.16
Weq log(MBH/M) 32 0.289 -0.33+0.65−0.43
Weq Eddington Ratio 32 0.168 -0.42+0.44−0.38
Weq αOX 31 0.002 -0.62+0.17−0.13
Weq log(LIR/LBOL) 32 0.067 0.37+0.20−0.25
Weq log(LFIR/LBOL) 30 0.094 -0.29+0.22−0.19
Weq log[N(H)/cm−2] 30 <0.001 0.19+0.03−0.03
Weq Γ 30 0.143 0.25+0.20−0.23
Weq log[F (0.5 − 2 keV)/erg s−1 cm−2] 26 0.005 -0.54+0.18−0.14
Weq log[F (2 − 10 keV)/erg s−1 cm−2] 29 0.002 -0.55+0.18−0.14
Weq log[L(0.5 − 2 keV)/erg s−1] 26 0.072 -0.33+0.22−0.19
Weq log[L(2 − 10 keV)/erg s−1] 30 0.009 -0.51+0.20−0.15
Weq log[L(0.5 − 10 keV)/erg s−1] 26 0.051 -0.38+0.23−0.18
Weq log[L(0.5 − 2 keV)/L(0.5 − 10 keV)] 26 0.192 0.21+0.22−0.24
Weq log[L(0.5 − 10 keV)/LBOL] 26 0.211 -0.20+0.25−0.21
vwtavg log(LBOL/L) 20 0.138 -0.27+0.25−0.22
vwtavg log[λL1125/erg s−1] 20 0.283 -0.15+0.25−0.22
vwtavg αOX 20 0.103 0.31+0.21−0.24
vwtavg log(LIR/LBOL) 20 0.115 0.37+0.25−0.30
vwtavg log(LFIR/LBOL) 20 0.496 0.00+0.25−0.26
vwtavg AGN fraction 20 0.340 0.35+0.49−0.80
vwtavg log(LAGN/L) 20 0.150 -0.24+0.23−0.21
vwtavg log(MBH/M) 20 0.286 -0.31+0.56−0.42
vwtavg Eddington Ratio 20 0.414 0.13+0.53−0.59
vwtavg log[N(H)/cm−2] 18 0.034 -0.15+0.08−0.08
vwtavg Γ 18 0.011 0.67+0.14−0.22
vwtavg log[F (0.5 − 2 keV)/erg s−1 cm−2] 17 0.008 0.61+0.15−0.21
vwtavg log[F (2 − 10 keV)/erg s−1 cm−2] 17 0.016 0.57+0.16−0.23
vwtavg log[L(0.5 − 2 keV)/erg s−1] 17 0.087 0.36+0.22−0.26
vwtavg log[L(2 − 10 keV)/erg s−1] 18 0.264 0.17+0.25−0.28
vwtavg log[L(0.5 − 10 keV)/erg s−1] 17 0.211 0.23+0.25−0.28
vwtavg log[L(0.5 − 2 keV)/L(0.5 − 10 keV)] 17 0.006 0.69+0.13−0.20
vwtavg log[L(0.5 − 10 keV)/LBOL] 17 0.009 0.64+0.14−0.20
σrms log(LBOL/L) 20 0.238 0.18+0.23−0.25
σrms log[λL1125/erg s−1] 20 0.425 -0.04+0.24−0.24
σrms αOX 20 0.004 -0.55+0.20−0.15
σrms log(LIR/LBOL) 20 0.305 -0.16+0.32−0.30
σrms log(LFIR/LBOL) 20 0.433 -0.04+0.25−0.25
σrms AGN fraction 20 0.389 -0.13+0.55−0.58
σrms log(LAGN/L) 20 0.247 0.17+0.23−0.25
σrms log(MBH/M) 20 0.301 0.29+0.43−0.59
σrms Eddington Ratio 20 0.374 -0.18+0.58−0.49
σrms log[N(H)/cm−2] 18 0.086 0.10+0.08−0.07
σrms Γ 18 0.048 -0.46+0.26−0.20
σrms log[F (0.5 − 2 keV)/erg s−1 cm−2] 17 0.016 -0.56+0.22−0.16
σrms log[F (2 − 10 keV)/erg s−1 cm−2] 17 0.015 -0.55+0.22−0.16
σrms log[L(0.5 − 2 keV)/erg s−1] 17 0.066 -0.40+0.26−0.21
σrms log[L(2 − 10 keV)/erg s−1] 18 0.137 -0.28+0.26−0.23
σrms log[L(0.5 − 10 keV)/erg s−1] 17 0.120 -0.32+0.27−0.23
σrms log[L(0.5 − 2 keV)/L(0.5 − 10 keV)] 17 0.022 -0.54+0.24−0.18
σrms log[L(0.5 − 10 keV)/LBOL] 17 0.008 -0.61+0.21−0.15
Note—Column (1): Dependent variable (absorption line property). Column (2):Independent variable (quasar/host property). Column (3): Number of points.Column (4): p-value of null hypothesis (no correlation). Column (5): Correlationcoefficient and 1σ errors. Underlined entries under col. (2) indicate significantcorrelations with p-values below 0.05.
rate of incidence of outflows among quasars with a steepX-ray to optical spectral index (αOX < −1.6; 75%) isalso higher than those with a shallow index (55%), al-though the Fisher exact test shows that this differenceis not significant (p = 0.45). A similar dependence onthe X-ray properties of the quasars has been reported inseveral studies of higher luminosity quasars and lowerluminosity Seyfert 1 galaxies using C IV λλ1548, 1550as a tracer of warm ionized outflows. We return to thisresult in Sections 6.3 and 6.4, and 7.
6.2. Optical Depths and Covering Factors
The distributions of the N V λ1243 and O VI λ1038optical depths and covering factors derived from theindividual components in the multi-component fits arepresented as histograms in Figure 6.
<0.1 0.1−0.4 0.4−1.6 1.6−50.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7Log(τ1243 or τ1038) Distributions
<0.1 0.1−0.4 0.4−1.6 1.6−50.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7 NVOVINV+OVI
NVOVINV+OVI
0.0−0.2 0.2−0.4 0.4−0.6 0.6−0.8 0.8−1.00.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7Covering Factor Distribution
0.0−0.2 0.2−0.4 0.4−0.6 0.6−0.8 0.8−1.00.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7 NVOVINV+OVI
NVOVINV+OVI
Figure 6. Distributions of the optical depths (left) andcovering factors (right) of the individual components used tofit the profiles of the N V (orange), O VI (purple), or jointN V + O VI (pink) absorption features.
Again, we repeat that the optical depths and coveringfactors presented here are only approximations. Never-theless, it is clear from the left panel in Figure 6 thata significant fraction of the absorbing systems are af-fected by saturation effects (τ1243 or τ1038 > 1), thereforemaking the equivalent widths of the N V and O VI fea-tures unreliable indicators of the total column densitiesof highly ionized gas in many of these cases.The right panel of Figure 6 shows that the mode of
the distribution of covering factors is consistent withunity, but ∼50% of the N V and O VI absorbers onlypartially cover the FUV quasar continuum emission (+possibly the broad emission line region − BELR; Fig. 6),consistent with small clouds located relatively near thequasars. As described at the end of Sec. 5.1, emissioninfill of the absorption profiles associated with the broadwings of the COS LSF is negligible and thus does notaffect this conclusion. We return to this result in Section7.1.
6.3. Outflow Equivalent Widths
Highly Ionized N V and O VI Outflows in the QUEST Quasars 17
The velocity-integrated equivalent widths (Weq; eq. 2)of the outflow systems in each quasar are listed in Table3. They span a broad range from ∼ 25 Å down to 20mÅ, near our 3-σ detection limit.The equivalent widths of the outflows were compared
against the properties of the quasars and host galaxieslisted in Table 1. Some of the results are shown in Figure7. By and large, we do not find any significant trendsbetween Weq and any of the quasar and host properties,except with some of the quantities that are derived fromthe X-ray data (Table 5). Taken at face value, this re-sult is surprising since, for instance, it means that theequivalent width of the outflow is largely agnostic of theproperties of the central engine over a range of ∼ 1.5 dexin power (FUV, bolometric, or quasar-only luminosity),∼2.0 dex in Eddington ratio, and ∼2.5 dex in black holemass. The lack of correlations with the properties of thehosts is less surprising since the quasar sample spans arelatively narrow range of values in these quantities sothe lack of correlation with these quantities may be at-tributed to the lower dynamical range.Examples of trends between Weq and the X-ray prop-
erties of the quasars are shown in panels (e), (f), and (g)of Figure 7. In panel (e), the equivalent width of the out-flow decreases with increasing HX luminosity. Panel (f)in this figure illustrates the dependence of the rate of in-cidence of these outflows on the X-ray column densitiesalready pointed out in Section 6.1. The stronger highlyionized outflows with Weq & 1 Å are only present inquasars with X-ray column densities above ∼ 1022 cm−2.While it is a required condition for a strong outflow, it isnot a sufficient condition since most quasars with theseX-ray absorbing column densities show either weak out-flows in the FUV (Weq < 0.3 Å) or none at all. Panel(g) also shows a distinct trend for strong outflows withWeq & 1 Å among objects with αOX . −1.7. A similartrend is observed when normalizing the X-ray luminosi-ties to the bolometric luminosities (not shown), but dis-appears when considering only the X-ray slope (e.g. theSX/HX ratio or index of the best-fit absorbed power-lawdistribution to the X-rays; not shown). Similar resultshave been found when considering C IV outflows (e.g.Brandt et al. 2000; Laor & Brandt 2002; Baskin & Laor2005; Gibson et al. 2009a,b). We return to this issue inSection 7 below.
6.4. Outflow Kinematics
Figure 8 shows the distributions of the velocity cen-troids and dispersions (σ) of the various individual com-ponents that were used to fit the N V and O VI absorbersin the quasar sample. About half of all of the indi-vidual components have blueshifted (outflow) velocities
ï2.0 ï1.6 ï1.2 ï0.8 ï0.4 0.0 0.4ï2
ï1
0
1
2
ï2.0 ï1.6 ï1.2 ï0.8 ï0.4 0.0 0.4Eddington Ratio
ï2
ï1
0
1
2
log(
Weq
/ Å)
p=0.168 r=ï0.42ï0.38+0.44 N=32
11.5 12.0 12.5 13.0-2
-1
0
1
2
11.5 12.0 12.5 13.0log(Lbol/LO •)
-2
-1
0
1
2
log(
Weq
/ Å)
p=0.046 r=ï0.34ï0.17+0.19 N=32
20 21 22 23 24ï2
ï1
0
1
2
20 21 22 23 24log[ N(H, Xïray) / cmï2]
ï2
ï1
0
1
2
log(
Weq
/ Å)
p<0.001 r=0.19ï0.03+0.03 N=30
ï2.2 ï2.0 ï1.8 ï1.6 ï1.4 ï1.2ï2
ï1
0
1
2
ï2.2 ï2.0 ï1.8 ï1.6 ï1.4 ï1.2_ox
ï2
ï1
0
1
2
log(
Weq
/ Å)
p=0.002 r=ï0.62ï0.13+0.17 N=31
ï0.6 ï0.4 ï0.2 0.0ï2
ï1
0
1
2
ï0.6 ï0.4 ï0.2 0.0log(LIR/Lbol)
ï2
ï1
0
1
2
log(
Weq
/ Å)
p=0.067 r=0.37ï0.25+0.20 N=32
43 44 45 46ï2
ï1
0
1
2
43 44 45 46log[L(2ï10 keV)/erg sï1]
ï2
ï1
0
1
2
log(
Weq
/ Å)
p=0.009 r=ï0.51ï0.15+0.20 N=30
6.8 7.2 7.6 8.0 8.4 8.8 9.2-2
-1
0
1
2
6.8 7.2 7.6 8.0 8.4 8.8 9.2log(MBH/MO •)
-2
-1
0
1
2
log(
Weq
/ Å)
p=0.289 r=ï0.33ï0.43+0.65 N=32
0.7 0.8 0.9 1.0ï2
ï1
0
1
2
0.7 0.8 0.9 1.0AGN fraction
ï2
ï1
0
1
2
log(
Weq
/ Å)
p=0.094 r=0.78ï0.61+0.19 N=32(a) (b)
(c) (d)
(e) (f)
(g) (h)
Figure 7. The velocity-integrated equivalent widths, Weq,of the outflow systems in the QUEST quasars are plotted asa function of the (a) bolometric luminosities, (b) AGN bolo-metric fractions, (c) Eddington ratios, (d) black hole masses,(e) hard X-ray (2 − 10 keV) luminosities, (f) X-ray absorb-ing column densities, (g) X-ray to optical spectral indices,and (h) ratios of the infrared luminosities to the bolometricluminosities. Red squares mark N V λλ1238, 1243 and bluecircles mark O VI λλ1032, 1038. Triangles indicate upperlimits (in one or both quantities). The regression results (p-values, correlation coefficients r with 1σ errors, and numberof points N ; Section 5) are shown at the top of each panel.The actual points used in the regression, in which N V andO VI quantities and/or X-ray measurements are averagedfor a given quasar, are shown in each inset panel. The solidpoints are detections, while the open points are censored val-ues in one or both quantities plotted.
that lie between [−2000, 0] km s−1 and have 1-σ widths
18 Veilleux et al.
less than 40 km s−1. In a blindly selected sample of O VIabsorbers, Tripp et al. (2008) similarly found that themajority of associated absorbers are within 2000 km s−1
of the QSO redshift (see their Figure 15). Likewise, theyfound that the O VI line widths are < 40 km s−1. Upto ∼10% of the individual components in the presentsurvey have redshifted velocities of up to a few × 100km s−1; some of them may be attributed to uncertainor systematically blueshifted systemic velocities derivedfrom the quasar emission lines (Sec. 5) rather than ac-tual inflows.
−8000 −6000 −4000 −2000 > 00.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7Velocity (km/s) Distributions
−8000 −6000 −4000 −2000 > 00.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
to −6000 to −4000 to −2000 to 0
NVOVINV+OVI
NVOVINV+OVI
10−40 40−80 80−160 >1600.0
0.2
0.4
0.6
0.8Log[σ / km/s ] Distributions
10−40 40−80 80−160 >1600.0
0.2
0.4
0.6
0.8 NVOVINV+OVI
NVOVINV+OVI
Figure 8. Distributions of the median velocities (left) andvelocity dispersions (right) of the individual componentsused to fit the profiles of the N V (orange), O VI (purple),and joint N V + O VI (pink) absorption features.
More physically meaningful kinematic quantities arethe weighted average velocities and velocity dispersionsof the outflow systems in each object (eqs. 3 and 4). Ofthe 20 detected absorbers in Table 3, 8 (3) have weightedaverage outflow velocities (velocity dispersion) in excessof 2000 (1000) km s−1.We find in Table 5 that there is no distinct trend be-
tween the weighted outflow velocities and velocity dis-persions and the quasar and host properties, except forthe lack of outflows in X-ray unabsorbed quasars (Fig.10), pointed out in Sec. 6.1, and the larger weighted out-flow velocity dispersions among X-ray faint sources withαOX . −2. The lack of a correlation between outflowvelocities and the quasar luminosities seems at odds withthose from most previous C IV absorption-line studies(e.g. Perry & O’Dell 1978; Brandt et al. 2000; Laor &Brandt 2002; Ganguly et al. 2007; Ganguly & Broth-erton 2008; Gibson et al. 2009a,b; Zhang et al. 2014;Rankine et al. 2020) and other multi-wavelength analy-ses (e.g., references in Sec. 1 and Veilleux et al. 2020).We examine this issue in more detail in Section 7 below.
7. DISCUSSION
7.1. Origins of the Absorption Features
20 21 22 23 24
0
ï2000
ï4000
ï6000
20 21 22 23 24log[ N(H, Xïray) / cmï2]
0
ï2000
ï4000
ï6000
v wta
vg (k
m/s
)
p=0.034 r=ï0.15ï0.08+0.08 N=18
7.0 7.5 8.0 8.5 9.0
0
-2000
-4000
-6000
7.0 7.5 8.0 8.5 9.0log(MBH/MO •)
0
-2000
-4000
-6000
v wta
vg (k
m/s
)
p=0.286 r=-0.31-0.42+0.56 N=20
ï0.6 ï0.4 ï0.2 0.0
0
ï2000
ï4000
ï6000
ï0.6 ï0.4 ï0.2 0.0log(LIR/Lbol)
0
ï2000
ï4000
ï6000
v wta
vg (k
m/s
)
p=0.115 r=0.37ï0.30+0.25 N=20
ï2.0 ï1.8 ï1.6 ï1.4 ï1.2
0
ï2000
ï4000
ï6000
ï2.0 ï1.8 ï1.6 ï1.4 ï1.2_ox
0
ï2000
ï4000
ï6000
v wta
vg (k
m/s
)
p=0.103 r=0.31ï0.24+0.21 N=20
43 44 45 46
0
ï2000
ï4000
ï6000
43 44 45 46log[L(2ï10 keV)/erg sï1]
0
ï2000
ï4000
ï6000
v wta
vg (k
m/s
)
p=0.264 r=0.17ï0.28+0.25 N=18
ï1.5 ï1.0 ï0.5 0.0
0
ï2000
ï4000
ï6000
ï1.5 ï1.0 ï0.5 0.0Eddington Ratio
0
ï2000
ï4000
ï6000
v wta
vg (k
m/s
)
p=0.414 r=0.13ï0.59+0.53 N=20
11.5 12.0 12.5 13.0
0
-2000
-4000
-6000
11.5 12.0 12.5 13.0log(Lbol/LO •)
0
-2000
-4000
-6000
v wta
vg (k
m/s
)
p=0.138 r=-0.27-0.22+0.25 N=20
0.7 0.8 0.9 1.0
0
ï2000
ï4000
ï6000
0.7 0.8 0.9 1.0AGN fraction
0
ï2000
ï4000
ï6000
v wta
vg (k
m/s
)
p=0.340 r=0.35ï0.80+0.49 N=20(a) (b)
(c) (d)
(e) (f)
(g) (h)
Figure 9. Same as Fig. 7 but the weighted average veloci-ties.
The blueshifted N V and O VI absorption features re-ported in Section 6 may have several origins: quasar-driven outflows, starburst-driven winds, tidal debrisfrom the galaxy mergers, and intervening CGM. Herewe do not consider contamination of the quasar spec-tra by young stars since none of them show the obvi-ous spectral signatures of young stars (e.g., narrow andshallow N V or O VI absorption troughs accompaniedby redshifted emission). This is only an issue amongstarburst-ULIRGs (e.g. Martin et al. 2015).Telltale signs that the detected lines are formed in a
quasar-driven outflow include (1) line profiles that are
Highly Ionized N V and O VI Outflows in the QUEST Quasars 19
20 21 22 23 240
1000
2000
3000
20 21 22 23 24log[ N(H, Xïray) / cmï2]
0
1000
2000
3000
m rm
s (km
/s)
p=0.086 r=0.10ï0.07+0.08 N=18
ï0.6 ï0.4 ï0.2 0.00
1000
2000
3000
ï0.6 ï0.4 ï0.2 0.0log(LIR/Lbol)
0
1000
2000
3000
m rm
s (km
/s)
p=0.305 r=ï0.16ï0.30+0.32 N=20
ï2.0 ï1.8 ï1.6 ï1.4 ï1.20
1000
2000
3000
ï2.0 ï1.8 ï1.6 ï1.4 ï1.2_ox
0
1000
2000
3000
m rm
s (km
/s)
p=0.004 r=ï0.55ï0.15+0.20 N=20
43 44 45 460
1000
2000
3000
43 44 45 46log[L(2ï10 keV)/erg sï1]
0
1000
2000
3000
m rm
s (km
/s)
p=0.137 r=ï0.28ï0.23+0.26 N=18
ï1.5 ï1.0 ï0.5 0.00
1000
2000
3000
ï1.5 ï1.0 ï0.5 0.0Eddington Ratio
0
1000
2000
3000
m rm
s (km
/s)
p=0.374 r=ï0.18ï0.49+0.58 N=20
7.0 7.5 8.0 8.5 9.00
1000
2000
3000
7.0 7.5 8.0 8.5 9.0log(MBH/MO •)
0
1000
2000
3000m
rms (
km/s
)p=0.301 r=0.29ï0.59
+0.43 N=20
0.7 0.8 0.9 1.00
1000
2000
3000
0.7 0.8 0.9 1.0AGN fraction
0
1000
2000
3000
m rm
s (km
/s)
p=0.389 r=ï0.13ï0.58+0.55 N=20
11.5 12.0 12.5 13.00
1000
2000
3000
11.5 12.0 12.5 13.0log(Lbol/LO •)
0
1000
2000
3000
m rm
s (km
/s)
p=0.238 r=0.18ï0.25+0.23 N=20(a) (b)
(c) (d)
(e) (f)
(g) (h)
Figure 10. Same as Fig. 7 but for the weighted velocitydispersions.
blueshifted, broad, and smooth compared to the ther-mal line widths (. 10 − 20 km s−1 for highly ionizedN4+, P4+, and O5+ ions at T ' 104.5−5.5 K), (2) lineratios within the multiplets N V λ1238/λ1243, O VIλ1032/λ1038, and P V λ1117/λ1128 that imply partialcovering of the quasar emission source, and (3) and largecolumn densities in these high-ionization ions (Hamannet al. 1997b,a; Tripp et al. 2008; Hamann et al. 2019b).N V is typically very weak or absent in intervening sys-tems (Werk et al. 2016). High N V/H I and O VI/H Iare also much higher in associated absorbers than in in-tervening systems (e.g., Tripp et al. 2008).
Among the 20 quasars with N V or O VI absorp-tion systems suggestive of outflows, 17 objects haveabsorption line profiles that meet the first of theabove criteria (the only exceptions are PG 0804+761,PG 0844+349, and PG 2233+134). Many of the quasarswith blueshifted N V or O VI absorption lines show N Vλ1238/λ1243 and/or O VI λ1032/λ1038 line ratios thatalso meet criteria #2 and #3 (Fig. 6). Mrk 231 doesnot formally meet these criteria (since it has no N Vabsorption line and O VI falls outside of the spectralrange of the data), but it shows all of the characteristicsof a FeLoBAL at visible and NUV wavelengths (and itsLyα line emission is highly blueshifted; Veilleux et al.2013b, 2016, and references therein), so we include ithere among those with quasar-driven outflows. So, over-all, at least 18 quasars in our sample have absorptionfeatures suggestive of quasar-driven outflows.In 15 of the 20 absorber detections, the velocity
widths, FWHMrms ≡ 2.355 σrms, are below the mini-mum of 2000 km s−1 generally used for BALs (Weymannet al. 1981, 1991; Hamann & Sabra 2004; Gibson et al.2009a,b), so they fall in the category of mini-BALs (500< FWHMrms < 2000 km s−1) or NALs (FWHMrms <
500 km s−1). Moreover, in many cases, the profiles arehighly structured rather than smooth, and thus do notmeet the “BAL-nicity” criterion to be true BALs.The weak and narrow redshifted absorption features
in PG 0804+761 and PG 0844+349 are good candidatesfor infalling tidal debris.
7.2. Location and Structure of the Mini-BALs7.2.1. Depths of the Absorption Profiles
The depths of the mini-BALs may be used to putconstraints on the location of the outflowing absorbers.The source of the FUV continuum in these quasars ispresumed to be the accretion disk on scale of ∼ few× 1015 cm (. 0.01 pc), where we used equation (6)in Hamann et al. (2019a) assuming an Eddington ratioηEdd = 0.1. But it is clear from the spectra that inmany cases (e.g., PG 1001+054, 1004+130, 1126−041,1309+355, 1351+640, 1411+442, 2214+139) the mini-BALs absorb not only the FUV continuum emission butalso a significant fraction of the Lyα, N V, and O VIline emission produced in the BELR. The gas producingthe mini-BALs must therefore be located outside of theBELR on scales larger than
rBELR = 0.1
(LAGN
1046 erg s−1
)1/2
pc (5)
(e.g. Kaspi et al. 2005, 2007; Bentz et al. 2013). Theradius of the outer boundary of the BELR, rout, is likelyset by dust sublimation (Netzer & Laor 1993; Baskin &
20 Veilleux et al.
Laor 2018). For gas densities of 105 − 1010 cm−3, Baskin& Laor (2018) derive
rout' rsubl ' (3− 6)rBELR (6)
= (0.3− 0.6)
(LAGN
1046 erg s−1
)1/2
pc, (7)
where graphite grains of size ∼ 0.05 µm is assumed (Fig.5 in Baskin & Laor 2018). Note that these values aresmaller than those in Barvainis (1987) and Veilleux et al.(2020), which are based on silicate grains and lower gasdensities (thus lower evaporation temperatures).More can be said about the structure of the absorbing
gas from the fact that the N V and O VI absorptionfeatures are optically thick (τ & 1 − 5; Fig. 6) but arenot completely dark. The covering factors derived fromthe multi-component fits to the N V and O VI mini-BALs range from 0.1 to 1, a direct indication that theabsorbing material is compact and spatially inhomone-neous. The often structured velocity profiles of the N Vand O VI mini-BALs (N V in PG 1411+442 is arguablythe only exception) also suggest a high level of kinematicsub-structures in the outflows, different from the smoothBALs observed in high-luminosity quasars. These prop-erties of the mini-BALs may indicate one of two things:(1) Our line of sight is not aligned along the directionof the outflowing stream of gas as in the case for theBALs (e.g. Murray et al. 1995; Elvis 2000; Ganguly et al.2001), but instead intercepts only a small fraction of thisstream and results in a covering factor of the backgroundemission that is highly dependent on the inhomogene-ity of the outflowing gas. (2) Another equally plausi-ble explanation is that (structured) mini-BALs form inmore sparse outflows or (in the unified outflow modeldiscussed for high-z quasars) in more sparse outflow re-gions, e.g., at higher latitudes above the disk.
7.2.2. Variability of the Absorption Profiles
Additional constraints on the location and structureof the BALs and mini-BALS in our sample may be ob-tained from profile variability. There is a vast literatureon this topic (e.g., Gibson et al. 2008; Hamann et al.2008; Gibson et al. 2010; Capellupo et al. 2012; FilizAk et al. 2012, 2013; Grier et al. 2015; He et al. 2019;Yi et al. 2019, and references therein). In our sample,four of the mini-BALs (PG 1001+054, PG 1126−041,PG 1411+442, and PG 2214+139) have been observedat two different epochs or more, and can therefore besearched for mini-BAL profile variations. The emer-gence of a dense [log nH(cm−3) & 7] new outflowabsorption-line system in PG 1411+442 was reportedin Hamann et al. (2019b) and the detailed inferrence ofa distance . 0.4 pc from the quasar is not repeated here.
We present the archival COS spectra for the other threeobjects in Figure 11, normalized to the same FUV con-tinuum level to emphasize absorption profile variations.In PG 1001+054 (Fig. 11a), the dramatic (72%) de-
crease in the FUV continuum emission from June 2014to March 2019 is accompanied by a strengthening of thebroad Lyα, N V, and O VI emission lines in terms ofequivalent widths but no obvious change in the mini-BAL profiles. In PG 1126−041 (Fig. 11b), the moremodest (20%) decrease of the continuum emission fromApril 2012 to June 2014 are not accompanied by anyobvious variations in the equivalent widths of any of thebroad emission and absorption lines except for the mostblueshifted N V absorption features below −3000 km s−1
which show variations on timescales perhaps as shortas 12 days. The broad emission and absorption linesin PG 2214+139 (Fig. 11c) show no variations betweenNovember 2011 and September 2021 despite a 26% in-crease in the strength of the FUV continuum emission.The fast 12-day variability of the high-velocity N V
mini-BAL in PG 1126−041 may be interpreted in twodifferent ways. One possibility is that transverse mo-tions of the outflowing clouds across our line of sightto the continuum source and BELR are responsiblefor these changes (as in PG 1411+442; Hamann et al.2019b). A variant on this idea is that the changes inprofiles are due instead to the dissolution and creationof the absorbing clouds/clumps in the outflow as theytransit in front of the continuum source. In the otherscenario, changes in the ionization structure of the ab-sorbing clouds due to changes in the incident quasar fluxcause the absorbing N V and O VI columns to vary andreproduce the observations. If this is the case, the vari-ability timescale sets a constraint on the ionization or re-combination timescale, which depends solely on the inci-dent ionizing continuum and gas density (∼ 105 yrs/nH ,where nH is the number density of the clouds in cm−3;e.g., He et al. 2019).This last scenario predicts that changes in the FUV
continuum of the quasar will produce changes in themini-BAL. While changes are indeed observed in boththe FUV continuum emission and high-velocity N Vmini-BAL of PG 1126−041, the amplitudes of thesechanges are not correlated. From 2012-04-15 to 2014-06-01, the continuum emission strengthened while the N Vmini-BAL weakened. From 2014-06-01 to 2014-06-12,both the continuum emission and N V mini-BAL weak-ened. Finally, from 2014-06-12 to 2014-06-28, the con-tinuum emission remained constant to within 1% but theN V mini-BAL strengthened slightly. This lack of a di-rect connection between variations in the continuum andthe N V mini-BAL seems to disfavor the scenario where
Highly Ionized N V and O VI Outflows in the QUEST Quasars 21
(b) PG 1126-041(a) PG 1001+054
(c) PG 2214+139
Figure 11. Multi-epoch comparisons of the mini-BALs in (a) PG 1001+054, (b) PG 1126−041, and (c) PG 2214+139. Allspectra are normalized to match the continuum level blueward of Lyα or redward of N V and O VI. The multiplicative factoris indicated in the caption.
the mini-BAL variations are associated with changes inthe ionization structure of the absorbing clouds, unlesslog nH(cm−3) . 5-6 in which case r >> rout and timedelays associated with the finite recombination timescalecould be at play (cf. Hamann et al. 2019b).While more detailed modeling of the mini-BAL of
PG 1126−041 is beyond the scope of the present paper,the fact that the mini-BAL variations are only observedin N V and only at high velocities may be an opticaldepth effect: the cloud complex that produces the Lyαand O VI mini-BALs and low-velocity N V mini-BALmay be so optically thick to be immune to variations inthe ionizing continuum or tangential movement of the
absorbing gas across the continuum source (we returnto this topic in Sec. 7.4 below).
7.3. Driving Mechanisms of the Mini-BALs
As reviewed in, for instance, Veilleux et al. (2020),the absorbing clouds making up the mini-BALs may bematerial (1) entrained in a hot, fast-moving fluid, or (2)pushed outward by radiation or cosmic ray pressure, or(3) created in-situ from the hot wind material itself. Inthe first two scenarios, the equation of motion of theoutflowing absorbers of massMabs that subtends a solidangle Ωabs is
d
dt[Mabs(r) r] = Ωabs r
2 (Pth + PCR + Pjet)
22 Veilleux et al.
+
(Ωabs4π
)(τLbol
c
)− GM(r)Mabs(r)
r2, (8)
where M(r) is the galaxy mass enclosed within a radiusr and τ is a volume- and frequency-integrated opticaldepth that takes into account both single- and multiple-scattering processes in cases of highly optically thickclouds (Hopkins et al. 2014, 2020).2 The terms on theright in Equation 8 are the forces due to the thermal,cosmic ray, and jet ram pressures, the radiation pressure,and gravity, respectively. Magneto-hydrodynamical ef-fects are assumed to be negligible at the distances ofthese absorbing clouds. The quasars in our sample donot have powerful radio jets so the jet ram pressure termcan safely be neglected. Similarly, the relatively mod-est radio luminosities of the mini-BAL quasars relativeto their optical and bolometric luminosities (Column 6in Table 1) suggest that cosmic-ray electrons do notplay an important dynamical role in accelerating theBAL clouds. Indeed the fraction of BAL quasars seemsto vary inversely with the radio loudness parameter, R(Column 5 in Table 1; e.g., Becker et al. 2001; Shankaret al. 2008). Below, we consider the remaining thermaland radiation pressure terms separately. In reality, thesepressure forces may act together to drive the mini-BALoutflows (see Sec. 7.4 for a closer look at the mini-BALPG 1126−041 in this context).
7.3.1. Thermal Wind and Blast Wave
For many years, ram-pressure acceleration of pre-existing clouds has been considered a serious contenderto explain BALs in quasars given the need for a muchhotter, rarefied medium to confine the clouds as they arebeing accelerated (e.g. Weymann et al. 1985). However,it is notoriouly difficult to accelerate dense gas cloudsfrom rest up to the typical (mini-)BAL velocities by awarm, fast thermal wind without destroying them inthe process through Rayleigh-Taylor fragmentation andshear-driven Kelvin-Helmholtz instabilities (e.g. Cooperet al. 2009; Scannapieco & Brüggen 2015; Schneider &Robertson 2015, 2017). Radiative cooling and magneticfields may act to slow down cloud disruption (Marcoliniet al. 2005; Cooper et al. 2009; Banda-Barragán et al.2016, 2018, 2020; Grønnow et al. 2018). Radiative cool-ing of the warm mixed gas can not only prevent disrup-tion, but it may cause the cloud to grow in mass (e.g.,Gronke & Oh 2018, 2020; Girichidis et al. 2021), al-
2 More explicitly, τ ≡ (1 − e−τsingle )(1 + τeff,IR). The value ofτ therefore ranges from ∼ τsingle = τUV/optical << 1 in theoptically thin case to ∼ (1 + τeff,IR) & 1 in the infrared optically-thick limit (the effective infrared optical depth, τeff,IR, is alsosometimes called the “boost factor”; Veilleux et al. (2020)).
though there are caveats (Schneider et al. 2020). An al-ternative scenario is that the BAL and mini-BAL cloudsare created in-situ via thermal instabilities and conden-sation from the hot gas with a cooling time shorterthan its dynamical time (Efstathiou 2000; Silich et al.2003). This is the idea behind the “blast wave” simula-tions of Richings & Faucher-Giguère (2018a,b); see alsoWeymann et al. (1985); Zubovas & King (2012, 2014);Faucher-Giguère & Quataert (2012); Nims et al. (2015).In these simulations, a fast (presumably X-ray emit-
ting) AGN wind with outward radial velocity of 30,000km s−1 is injected in the central 1 pc and collides vio-lently with the host ISM. The resulting shocked windmaterial reaches a very high temperature (∼1010 K;Nims et al. 2015) that does not efficienty cool, butinstead propagates outward as an adiabatic (energy-driven) hot bubble. This expanding bubble sweeps upgas and drives an outer shock into the host ISM raisingits temperature to a few × 107 K (Nims et al. 2015). Ra-diative cooling of the shocked ISM eventually becomesimportant and the outflowing material reforms as coolneutral and molecular gas, but by that time, the out-flowing material has acquired a significant fraction of theinitial kinetic energy of the hot wind. These simulationspredict that the cooling radius, i.e. the radius at whichthe gas cools to 104 K, increases from 100 pc to 1 kpcfor AGN with luminosities from 1044 to 1047 erg s−1, re-spectively (Fig. 7 of Richings & Faucher-Giguère 2018b).This cooling radius is also the expected location of thegas clouds producing the N V and O VI mini-BALs, asthe cooling gas rapidly transitions from ∼ 107 K to ∼104 K. This large radius is well outside the lower limit onthe distance of the mini-BAL from the quasars derivedabove so it is not inconsistent with our data. However,one should note that the inner X-ray wind in quasarsis presumed to much smaller in reality than the valueof 1 pc assumed in the simulations so the cooling ra-dius may have to be scaled down accordingly. Moreover,these models do not address how the BELRs would berestored after the passage of the blast wave. Finally,the detailed analysis of the BAL in PG 1411+442 firmlyrule out (at least in that case) absorption at these largedistances.
7.3.2. Radiative Acceleration
Although ram-pressure acceleration has been a seriouscontender, overall the favored explanation for the largevelocities of the BALs and mini-BALs is that the gas ab-sorbers have been accelerated by the radiation pressureforces associated with the intense radiation field thatis emanating from the quasars (e.g., Arav et al. 1994;Giustini & Proga 2019). Strong support for this sce-
Highly Ionized N V and O VI Outflows in the QUEST Quasars 23
nario comes from the observed trends for the maximumvelocity of the absorption to increase on average with in-creasing optical, UV, or bolometric luminosity and theEddington ratio (e.g., Perry & O’Dell 1978; Brandt et al.2000; Laor & Brandt 2002; Ganguly et al. 2007; Gan-guly & Brotherton 2008; Gibson et al. 2009b; Zhanget al. 2014). Note, however, that the overall correla-tions noted in these studies are often quite modest andsometimes only visible when considering the upper enve-lope of the velocity distribution and only when the sam-ple of AGN span 2-3 orders of magnitude in luminosity(sometimes combining NALs, mini-BALs, and BALs to-gether). While more recent studies (e.g., Rankine et al.2020) have confirmed and indeed strengthened the exis-tence of some of these correlations, all of the cited resultsrelate to the C IV absorption, rather than the N V andO VI features. The statistics on N V and particularlyO VI absorbers are much poorer.Far Ultraviolet Spectroscopic Explorer (FUSE) obser-
vations of Seyfert galaxies of relatively low luminosities(1038 − 1042 erg s−1) show either no or very weak trendsof increasing maximum velocities with increasing lumi-nosities and no trend at all with the Eddington ratio(Kriss 2004a,b; Dunn et al. 2008). O VI and N V BALsin high-redshift, high-luminosity quasars (Baskin et al.2013; Hamann et al. 2019a) have maximum velocitiesthat correlate with their C IV counterparts, but therange in AGN luminosity of their sample is too small totest the luminosity dependence of the maximum veloci-ties. More fundamentally, there is also a trend betweenline widths and optical depth. The most extreme exam-ple of this trend is P V, which coexists with C IV havingthe same ionization requirements, but is always weakerand narrower than C IV (Hamann et al. 2019a). Thereason is that P V traces only the highest column den-sity regions with smaller covering fractions, while C IVcan have significant absorption in more diffuse gas occu-pying a larger volume. This evidence for optical depth-dependent covering factors is a signature of inhomoge-neous partial covering.Overall, given the complex results from these previous
studies, it is perhaps not surprising to find no significantcorrelations in our sample of QUEST quasars between(maximum) outflow velocities and the AGN luminosities(Sec. 6.4). Theoretically, the noise in the trends betweenthe outflow kinematics and AGN luminosity is expectedin the radiative acceleration scenario given projectioneffects that reduce the measured outflow velocities andvariance in both the (minimum) launching radius (e.g.Laor & Brandt 2002) and efficiency of radiative acceler-ation associated with the complex micro-physics of thephoton interaction with the clouds - this complexity is
hidden in the quantity τ in equation 8. A similar trendof increasing variance in the maximum velocity with in-creasing AGN luminosity is observed in the other coolergas phases of AGN-driven outflows (e.g., Veilleux et al.2020; Fluetsch et al. 2020).Additional evidence that radiation pressure plays an
important role in accelerating the absorbers in quasarscomes from the significant dependence of the incidencerate, equivalent width, and weighted outflow velocitydispersion of the blueshifted absorbers on the X-rayproperties of the quasars. This effect has been reportedin numerous studies of nearby and distant AGN, basedlargely on C IV and Si IV (e.g., Laor & Brandt 2002;Gibson et al. 2009b), and is also clearly present in oursample of quasars based on N V and O VI (Sec. 6.1, 6.3,and 6.4). More specifically, we find that mini-BALs andBALs are broader, stronger, and more common amongX-ray faint quasars with steep optical-to-X-ray slopesαOX . −1.7 and hydrogen column densities NH in ex-cess of ∼ 1022 cm−2 (Sec. 6). This result is expectedin the context of radiative acceleration since the com-bined radiative force (“force multiplier”; Arav & Li 1994)is greatly reduced when the gas is over-ionized by thehard far-UV/X-rays, becoming too transparent to beradiatively accelerated. This over-ionized “failed-wind”material may act as a radiative shield to produce thespectral softening needed for efficient radiative accelera-tion of the outflow material downstream (Murray et al.1995; Proga & Kallman 2004; Proga 2007; Sim et al.2010). However, the strong near-UV absorption linesnear systemic velocity expected in this scenario are notobserved (Hamann et al. 2013). Alternatively, the spec-trum emerging from the accretion disk may be intrinsi-cally softer/fainter in the hard far-UV/X-rays than com-monly assumed (e.g. Laor & Davis 2014). Weak-lined“wind-dominated” quasars, such as Mrk 231, PHL 1811and its analogs, which are intrinsically faint and unab-sorbed in the X-rays, may be naturally explained in thisfashion (Richards et al. 2011; Wu et al. 2011; Luo et al.2015; Veilleux et al. 2016). While a connection shouldexist between the X-ray warm absorbers and the UVabsorption-line outflows, a direct one-to-one kinematiccorrespondence between the two classes of absorbers isoften not seen because the gas in the warm absorbersis too highly ionized to produce measurable lines in theUV spectra (Kaspi et al. 2000, 2001; Gabel et al. 2003;Kraemer et al. 2001; Krongold et al. 2003; Arav et al.2015; Laha et al. 2021, and references therein). We re-turn to this point in Section 7.4 when discussing themini-BAL in PG 1126−041 (the case of PG 1211+143is briefly discussed in Appendix A).
24 Veilleux et al.
Another observational characteristic of outflows thatfavors radiative acceleration is the phenomenon of line-locking observed in perhaps as many as ∼ 2/3 of allNAL and (mini-)BAL outflows (e.g., Hamann et al.2011; Bowler et al. 2014; Lu & Lin 2018; Mas-Ribas& Mauland 2019, and references therein). This is ob-served in outflows where multiple absorbers are presentbut are separated by the exact same velocity separa-tion as the doublet of C IV (499 km s−1), N V (962km s−1), or O VI (1650 km s−1). Previous studies haveshown that the probability of a line-locking signature ac-cidentally occurring over a relatively small redshift pathis negligible (e.g., Foltz et al. 1987; Srianand 2000; Sri-anand et al. 2002; Ganguly et al. 2003; Benn et al. 2005).Radiative acceleration is a natural explanation for line-locking (e.g., Mushotzky et al. 1972; Scargle 1973; Braun& Milgrom 1989).The best case for line-locking in our data is that of
PG 1351+640, where two deep absorption throughs aredetected in Lyα, extending over −[2200, 1500] km s−1
and −[1400, 600] km s−1, roughly separated by the ve-locity split of the N V doublet lines (∆v ≈ 900 − 1000km s−1 ≈ ∆vNV ; Fig. 12). This results in a N V mini-BAL that looks like a “triplet” in this object. Anothercase of line-locking may be present in PG 1126−041,where some of the deepest Lyα throughs are separatedby ∆v ≈ 900 − 1000 km s−1 ≈ ∆vNV (Fig. 3).Note, finally, that radiation pressure on dust grains
has also been invoked as an important contributor to theradiative acceleration given that (mini-)BAL QSOs, par-ticularly LoBALs and FeLoBALs, have more reddenedUV spectra than non-BAL QSOs (e.g., Allen et al. 2011;Hamann et al. 2019a, and references therein). This re-sult has been interpreted to mean that BAL and mini-BAL clouds have large columns of ionized + neutral gas(log NH(cm−2) & 23) and enough dust to provide ex-tinction equivalent to at least AV ∼ 1-2 mag. in somecases (this is discussed in Sec. 7.4 in the context ofPG 1126−041, but see also the results on Mrk 231 andPG 1411+442; Veilleux et al. 2016; Hamann et al. 2019b,respectively). Under these circumstances, the (mini-)BAL clouds may be subject to larger radiative forcesthan dustless clouds since the dust cross section, andthus τ in equation 8, in the UV optical is > 1-2 or-ders of magnitude than the Thompson scattering crosssection of electrons. Unfortunately, we do not have a re-liable reddening indicator of the FUV continuum emis-sion in our quasars, so we cannot directly compare ourdata with those of BAL and non-BAL QSOs. On theother hand, radiation that will be absorbed by the dustin the broad absorption line regions (BALRs) will bere-emitted in the infrared so our measurements of the
PG1351+640, z=0.088
−4000 −3000 −2000 −1000 0Velocity (km/s)
0.00.2
0.4
0.6
0.8
1.0
Nor
mal
ized
Fλ
NV 1238
0.00.2
0.4
0.6
0.8
1.0
NV 1242
0.00.2
0.4
0.6
0.8
1.0
Lyα
Figure 12. Line-locking in PG 1351+640. The two deepLyα absorption throughs are separated by the velocity splitof the N V doublet lines, resulting in a N V complex thatlooks like a “triplet”.
infrared excess in our quasars may serve as a surrogatefor the amount of dust in the BALRs. The lack of ob-vious correlation between BAL properties in our quasarsample and the mid-, far-, and total (1 − 1000 µm) in-frared excesses (e.g., Figs. 7 and 10) indicates one of twothings: (1) radiative acceleration on dust is not impor-tant in the BALRs of these objects or (2) the variousinfrared excesses are dominated by dust emission fromoutside of the BALR, e.g. dust in the host galaxy itself.
7.4. P V Mini-BAL in PG 1126−041 and OtherQuasars
In this last section, we take a closer look at the mini-BAL system in PG 1126−041. A mini-BAL extendingfrom −1000 to −5000 km s−1 was first reported in theN V, C IV, and S IV absorption lines of this objectby Wang & Wang (1999); Wang et al. (1999), based onthe analysis of old low-resolution spectra obtained withIUE and the Goddard High-Resolution Spectrograph(GHRS) on HST. Variable and much faster (∼16,500km s−1) X-ray absorption has also been detected in thisobject (Wang & Wang 1999; Wang et al. 1999; Gius-
Highly Ionized N V and O VI Outflows in the QUEST Quasars 25
tini et al. 2011). Interestingly, this object is among theleast luminous AGN (log LBOL/L = 11.52) in our sam-ple, intermediate between quasars and typical Seyfert 1galaxies. Mini-BALs with outflow velocities of up to∼ 5000 km s−1 and widths (FWHMrms ≡ 2.355 σrms)> 1000 km s−1 are relatively rare in such low-luminositysystems (e.g., Kriss 2004a,b; Dunn et al. 2008; Crenshaw& Kraemer 2012). On the other hand, PG 1126−041is also the object in our sample with the steepest X-ray to optical index (αOX = −2.13, a virtual tie withPG 1001+054, which also harbors a mini-BAL) andis among those with the largest infrared excess (Table1), reinforcing the view expressed in Section 7.3.2 thatX-ray absorbed or intrinsically weak quasars are morelikely to host BALs and mini-BALs.Apart from the line-locking signatures found in the
N V mini-BAL of this object, which we argued in Sec. 7.3favors radiative driving, the most remarkable aspect ofthis mini-BAL is the detection of a narrow P V λλ1117,1128 cloud at a velocity of −2200 km s−1 (Fig. 3). Largeionized-gas column densities are needed to produce thisfeature given the low abundance of phosphorus relativeto hydrogen (log(P/H) = −5.54; Lodders 2003). Themulti-component fit of each line in the P V doublet re-quires two components with nearly identifical medianvelocities (−2196 and −2203 km s−1) but different ve-locity dispersions (24 and 74 km s−1), covering factors(0.37 and 0.10, respectively), and optical depths (0.7and 2.6, respectively). The total equivalent width ofthis doublet is 0.3 Å.Taken at face value, the results from the fits suggest
that the P V lines are only moderately optically thickand therefore more reliable indicators of the total ion-ized column densities of this cloud than the highly sat-urated N V and O VI features. This is supported bythe ∼2:1 intensity ratio of the P V lines. An opticaldepth of order unity in P V λ1128 implies an ionizedhydrogen column density log NH(cm−2) ≈ 22.3, assum-ing a solar P/H abundance ratio and ionization correc-tions based on detailed photoionization calculations fortypical BALs and mini-BALs (ionization parameters logU & −0.5; Hamann 1998; Leighly et al. 2011; Borguetet al. 2012, 2013; Baskin et al. 2014; Capellupo et al.2017; Moravec et al. 2017; Hamann et al. 2019a, andreferences therein). This column density is remarkablyconsistent with the expectations from radiation pressureconfined cloud models (Baskin et al. 2014).This value of the total column density may be used to
estimate the minimum kinematic energy of this outflow-ing cloud using (eq. 2 from Hamann et al. 2019a)
Ekin = 1.5× 1052(
Q0.15
)(NH
2×1022 cm−2
)(r
1 pc
)2
×(
v2200 km s−1
)2ergs, (9)
where Q is an approximate global outflow covering fac-tor based on the incidence of mini-BALs in the SDSSquasars (Trump et al. 2006; Knigge et al. 2008; Gibsonet al. 2009b; Allen et al. 2011) and r = 1 pc is a place-holder radial distance that we adopt for illustration pur-poses (it may underestimate the actual distance of theabsorbers from the source; see Sec. 7.2 and Arav et al.2020, and references therein). Following Hamann et al.(2019b), we estimate the time-averaged kinetic energyluminosity, Lkin, by dividing Ekin by a characteristicflow time, r/v ≈ 450 yr. This yields Lkin & 1 × 1042
erg s−1. In these units, LBOL = 1.3 × 1045 erg s−1
so Lkin/LBOL & 0.001. Taken at face value, this ratiois too small to significantly affect the evolution of thegalaxy host (e.g., Lkin & 0.005 LEdd is needed accordingto Hopkins & Elvis 2010), unless (1) r is severely un-derestimated or (2) the other clouds at lower and highervelocities involved in this mini-BAL contribute signifi-cantly to Lkin. This second possibility seems unlikelygiven the lack of P V detection in these clouds whichsuggests column densities log NH(cm−2) . 22.Next, we use the total column density of the P V
cloud in PG 1126−041 to estimate the time-averagedmomentum outflow rate of this cloud, p = 2Lkin/v ≈1 × 1034 dynes, and compare this value with the radi-ation pressure, LBOL/c = 4 × 1034 dynes. Given thatp/(LBOL/c) ∼ 0.2, radiation pressure can thus in princi-ple accelerate this cloud, although a contribution from athermally driven wind as detailed in Section 7.3.1 cannotbe formally ruled out.Finally, we apply Equation 9 and calculate Lkin/LBOL
for the other mini-BALs in our sample with solid andtentative P V detections. For PG 1411+442, the onlyother mini-BAL in the sample with a definite P V detec-tion, we get Lkin/LBOL & 0.01 and p/(LBOL/c) ∼ 1, foran outflow velocity of −1800 km s−1 (Table 3), a totalcolumn density log NH(cm−2) & 23.4, and a BELR-likedistance . 0.4 pc from the central light source derivedby Hamann et al. (2019b) using several absorption linesand detailed photoionization simulations. This BALmay thus be sufficient to impact the host galaxy evo-lution. P V is also tentatively detected in PG 1001+054and PG 1004+130 at velocities of ∼ [−4000, −6,000]km s−1 (Fig. 13). In both cases, the equivalent widths ofP V 1117 and 1128 are very similar, implying saturationand log NH(cm−2) & 22.3. These numbers yield out-flows with kinetic-to-bolometric luminosity ratios thatare higher than that of PG 1126−041 but lower thanthat of PG 1411+442, and thus marginally sufficient toimpact the host galaxy evolution. Overall, the mini-BALs in the QUEST quasars are less powerful than the
26 Veilleux et al.
P V BALs detected in ze & 1.6 SDSS quasars (Moravecet al. 2017) but perhaps more common (4/33 ∼ 10%)than at high luminosities/redshifts (detection rate ofonly 3−6% among the ze & 2.6 BAL quasars in theSDSS-III BOSS quasar catalog Capellupo et al. 2017).
(a) PG 1001+054
(b) PG 1004+130
Figure 13. Tentative detection of P V in (a) PG 1001+054and (b) PG 1004+130. The spectra have been heavily binnedto emphasize the broad but shallow features. The deep andnarrow absorption lines in the velocity range −[8,000, 14,000]km s−1 in the P V 1117, 1128 panels of both PG 1001+054and PG 1004+130 are due to intervening MW ISM (Si II1259, 1260 + Fe II 1260 and C II 1334 + C II∗ 1335, respec-tively).
8. SUMMARY
As part I of a HST FUV spectroscopic study of theQUEST (Quasar/ULIRG Evolutionary Study) sampleof local quasars and ULIRGs, we have conducted a uni-form analysis of the COS spectra of 33 z . 0.3 Palomar-Green quasars. The main conclusions from our analysisare the following:
1. Highly ionized outflows traced by blueshifted N Vλλ1238, 1243 and O VI λλ1032, 1038 absorp-tion lines with equivalent widths larger than ∼20 mÅ are present in about 60% of the QUESTquasars. This detection rate is similar to thatof warm-ionized outflows traced by blueshiftedC IV λλ1548, 1550 absorption lines in localSeyfert galaxies and more distant, higher luminos-ity quasars.
2. The N V and O VI features in the QUEST quasarsspan a broad range of properties, both in termsof equivalent widths (from 20 mÅ to 25 Å) andkinematics (outflow velocities from a few × 100km s−1 up to ∼ 10,000 km s−1).
3. The rate of incidence and equivalent widths ofthe highly ionized outflows are higher among X-ray weak sources with X-ray to optical spectralindices αOX . −1.7 and X-ray column densitieslog NH(cm−2) & 22. The weighted outflow ve-locity dispersions are highest in the X-ray weak-est sources with X-ray to optical spectral indicesαOX . −2. These results are qualitatively similarto AGN-driven warm ionized outflows traced bythe C IV λλ1548, 1550 absorption lines. These re-sults favor radiative acceleration of the absorbers,where the X-rays are either absorbed or intrinsi-cally weak in the wind-dominated systems. Line-locking is detected in the Lyα absorption throughsof one or two objects, providing additional evi-dence that radiation pressure plays an importantrole in accelerating these absorbers.
4. There is no significant trend between the weightedaverage velocity of the highly ionized outflows andthe properties of the quasars and host galaxies.This negative result is likely due in part to thefact that the range of properties of the QUESTquasar sample is narrow in comparison to those ofother studies.
5. Blueshifted P V broad absorption lines are clearlydetected in PG 1126−041 and PG 1411+442 (pre-viously reported in Hamann et al. 2019b), and also
Highly Ionized N V and O VI Outflows in the QUEST Quasars 27
posssibly in PG 1001+054 and PG 1004+130. Us-ing the results from the analysis of Hamann et al.(2019b), these features imply column densities of∼ 1022.3 cm−2 or larger and time-averaged outflowkinetic power to bolometric luminosity ratios of &0.1% if a conservatively small radial distance of 1pc from the P V absorbers is assumed.
Paper II of this series (Liu et al. 2021, in prep.) willpresent the results from our analysis of the COS spectraon the QUEST ULIRGs. These results will be combinedwith those from the present paper to provide a morecomplete picture of the gaseous environments of quasarsand ULIRGs as a function of host galaxy properties andage across the merger sequence from ULIRGs to quasars.
ACKNOWLEDGMENTS
The authors thank the anonymous referee for sugges-tions which improved this paper. SV, WL, and TMTacknowledge partial support for this work provided byNASA through grant numbers HST GO-1256901A andGO-1256901B, GO-13460.001-A and GO-13460.001-B,and GO-15662.001-A and GO-15662.001-B from theSpace Telescope Science Institute, which is operatedby AURA, Inc., under NASA contract NAS 5-26555.Based on observations made with the NASA/ESA Hub-ble Space Telescope, and obtained from the HubbleLegacy Archive, which is a collaboration between theSpace Telescope Science Institute (STScI/NASA), theSpace Telescope European Coordinating Facility (ST-ECF/ESA) and the Canadian Astronomy Data Cen-tre (CADC/NRC/CSA). The authors also made use ofNASA’s Astrophysics Data System Abstract Service andthe NASA/IPAC Extragalactic Database (NED), whichis operated by the Jet Propulsion Laboratory, Califor-nia Institute of Technology, under contract with the Na-tional Aeronautics and Space Administration.
Software: COSQUEST (Rupke 2021a), IFSFIT(Rupke 2014; Rupke & Veilleux 2015; Rupke 2021b),LINMIX_ERR (Kelly 2007), pymccorrelation (Privonet al. 2020), scipy (Virtanen et al. 2020), DRTOOLS(Rupke 2021c)
28 Veilleux et al.
APPENDIX
A. DETAILED RESULTS FROM THE SPECTRAL ANALYSIS
Figures 14a−14t present the fits to the detected N V, O VI, and P V absorption systems in our sample. The resultsfrom these fits are listed in Table 3 in the main body of the paper. Here we summarize the results from our spectralanalysis for each object in the sample.
PG 0007+106.—There are no associated N V absorbers in this system, although two deep blueshifted and redshiftedLyα absorption features are present at |v| < 400 km s−1.PG 0026+129.—There are no associated N V or O VI absorbers in this object.PG 0050+124 (I Zw 1).—N V and Lyα absorbers are detected at −553, −1315, and −1467 km s−1 in this object.
Variable warm absorbers at −1870 and −2500 km s−1 have been reported by Silva et al. (2018) in XMM-Newton RGSspectra obtained in 2015.PG 0157+001 (Mrk 1014).—There are no associated N V or O VI absorbers in this system.PG 0804+761.—This is a rare case for infall where a strong redshifted absorption system at +600 km s−1 is observed
in Lyα, and a corresponding weaker feature is also detected in N V and O VI. The stronger features in the panel ofFigure 14b labeled O VI 1037 are Fe II features from the MW ISM.PG 0838+770.—A weak low-|v| absorption feature is seen in Lyα but not N V or O VI.PG 0844+349.—A strong slightly redshifted double-component absorption feature, likely associated with tidal debris,
is present at +140 km s−1 and +190 km s−1 in both Lyα and N V, but the feature at ∼ 1288.9 Å has no correspondingN V and is presumed to be Lyα from intervening CGM.PG 0923+201.—A single broad absorber is observed at −3300 km s−1 in Lyα, Lyβ, and O VI 1032 and 1038 although
the glare of the geocoronal Lyα airglow truncates the blue wing of the O VI 1032 absorption line.PG 0953+414.—Two faint O VI features are detected at −140 and −1074 km s−1. The first feature is also visible
in Lyα and Lyβ, but the more blueshifted feature is not. The strong feature near 1271 Å cannot be fit with O VI andthus is likely Lyα from intervening CGM.PG 1001+054.—The multi-epoch COS spectra shown in Figure 11 are co-added for this analysis given the lack of
variability in the absorption line profiles (Sec. 7.2). This is a special case because of the way the broad, high-velocityN V absorption absorbs the Lyα profile and the O VI absorption is deep, nearly dark, and highly saturated. The threemethods discussed in Section 5 were attempted to deal with the Lyα + N V blend, and in the end method #3 was usedfor the final fit: (1) The use of polynomials/splines to fit the blue and red sides of Lyα is problematic because it doesnot appear to yield a symmetric Lyα line and seems to miss N V absorption that appears in Lyα. (2) A Lorentzianfit to Lyα works reasonably well in that it yields a symmetric line, but the line properties are highly unconstrained(particularly in terms of height) and sensitive to the choice of which continuum regions are fit. (3) After trying theQSO templates from Vanden Berk et al. (2001), Stevans et al. (2014), and Harris et al. (2016), we settled on the lastone. We scaled the template using a constant offset, a constant multiplier, and a scaling according to λp (arbitrarypower p) to account for differences between the spectral index and that of the template (following Harris et al. 2016).We fit only the blue side of Lyα (and the far-red side) as well as some continuum regions in between O VI and LyαN Vthat are fairly line-free. This underpredicts the strength of highly-ionized lines, but is the best compromise solution.If the lines are fit as well, the blue side of Lyα is not properly fit. The resulting fit to N V seems to work reasonablywell, although the fit results are obviously illustrative in terms of velocity space and certainly do not get the opticaldepth correct. There seems to be a narrow Lyα near −4000 km s−1 that is barely detected in N V, but the fit hasdifficulties capturing it. The broad, highly saturated, nearly dark O VI is fit over −[6000, 4000] km s−1 but not −[8000,6000] km s−1 to account for Lyβ contamination. Broad blueshifted P V 1117, 1128 at ∼ −[6000, 4000] km s−1 is alsotentatively detected in this object but no attempt is made to fit this faint feature (Fig. 13a).PG 1004+130.—This is another special case (see Wills et al. 1999; Brandt et al. 2000, for some previous analyses).
It is difficult to fit the “continuum” shape and O VI profile simultaneously. The present fit is the best availablecompromise. It is clear that Lyβ / O VI are interacting with each other so the O VI fit should be taken only asillustrative. There are higher-order Lyman lines in the spectrum that are not considered. The Lyγ region is shown butis not considered in the fits to O VI (the many narrow features in the Lyγ region are Galactic ISM lines). BlueshiftedP V 1117 + 1128 at −[6000, 4000] km s−1 is tentatively detected but no attempt is made to fit this faint feature (Fig.
Highly Ionized N V and O VI Outflows in the QUEST Quasars 29
13b).PG 1116+215.—There are several narrow absorption features blueward of Lyα, O VI, and Lyβ in this system, but
only two of them are paired in O VI λ1032 and λ1038, corresponding to −771 and −2825 km s−1. Presumably, theorphan features are intervening Lyα absorbers. Savage et al. (2014) also report an O VI system at 1175 and 1181 Å,but it is not shown here as it is very narrow and very high velocity, and is intervening CGM (there are nearby galaxiesat the same redshift Tripp et al. 1998).PG 1126−041.—Several separate COS/G130M spectra exist for this object (Table 2). Our analysis uses the co-added
spectrum. Broad blueshifted absorption features are clearly detected in Lyα, N V, and O VI (Fig. 3). A narrow P Vfeature is observed at −2200 km s−1, indicative of large ionized column densities. This P V detection is discussed inmore detail in Section 7.4. 2014 and 2015 COS/G160M spectra centered around ∼ 1450 Å show strong C IV λλ1548,1550 absorption throughs similar to those of N V and O VI, but only weak and narrow blueshifted S IV λλ1392, 1402absorption features with |v| . 500 km s−1; the analysis of these features is beyond the scope of the present study.PG 1211+143.—There are no associated N V absorption systems in this object. A broad absorption feature at the
observed wavelength of 1240 Å has been interpreted by Kriss et al. (2018) as a highly blueshifted (−16,980 km s−1)and broad (FWHM ≈ 1080 km s−1) Lyα absorption from a fast wind. It may correspond to one of the ultra-fastoutflowing systems detected in the X-rays (Danehkar et al. 2018, and references therein). However, since it is notdetected in N V it is not included in the statistical analysis of Section 6.PG 1226+023 (3C 273).—The FUV spectrum of 3C 273 has been used extensively to study the low-z IGM (e.g.,
Tripp et al. 2008; Savage et al. 2014). There are no associated N V or O VI absorption systems in this object.PG 1229+204.—There are no associated N V absorption systems in this object. The many unidentified features
shortward of Lyα in the quasar rest-frame (e.g., 1289.5, 1282, 1223.2, 1220.4 Å) may be Lyα from intervening CGM.Mrk 231.—The 2011 COS/G130M spectrum of this object was presented in Veilleux et al. (2013a), while the
2014 COS/G140L and G230L spectra were presented in Veilleux et al. (2016). The COS/G130M spectrum shows Lyαemission that is broad (& 10,000 km s−1) and highly blueshifted (centroid at ∼ −3500 km s−1). In contrast, blueshiftedabsorption features are only present above ∼2200 Å. These results have been discussed in details in Veilleux et al.(2016), and this discussion is not repeated here. This outflow is considered a non-detection in our analysis in Section 6since it has no N V absorption systems, but has all of the characteristics of a FeLoBAL at visible and NUV wavelengthsand is considered as such in our discussion (Sec. 7).PG 1302−102.—There are no associated O VI absorbing systems in this object, but several intervening Lyα and
metal-line systems have been reported by Cooksey et al. (2008).PG 1307+085.—The weak O VI absorber at ∼ −3400 km s−1 is also detected in Lyα and Lyβ. The other feature
at ∼ −3600 km s−1 is seen only in Lyβ and Lyα and is presumed to be from intervening CGM.PG 1309+355.—A broad absorption feature is visible in Lyα, extending blueward to ∼ −1600 km s−1. This feature
is also detected in both O VI lines, but is truncated by the gap associated with the strong geocoronal Lyα emission.Strong absorption features are detected coincident with P V 1117 and 1128 near systemic velocity, but given theirequivalent widths they are likely of Galactic origin (e.g. C II 1334).PG 1351+640.—Two deep broad absorption throughs are detected in N V and Lyα of this object, extending over
[−2200, −1500] km s−1 and [−1400, −600] km s−1. They are roughly separated by the velocity separation of the N Vdoublet lines, resulting in a N V “triplet”, so this is a good case for line locking (Sec. 7.3.2). The template fit to the Lyαline emission works well in the very narrow region in which it is used, but mostly because it gets the smooth profile inthis region better than a spline+polynomial fit. The central peak is steeper than the template, and it is possible thatthe bluest absorption region in Lyα at −[3500, 2500] km s−1 is caused by a weak peak rather than a true absorption,since it does not line up with anything in N V. The N V template fit, again over this limited range, is quite good andmore trustworthy over a wider wavelength range. It is actually quite comparable to the spline fit. It does not get thesteep central peak in Lyα on the red side.PG 1411+442.—This object has been the subject of a detailed analysis by Hamann et al. (2019b). A deep broad
absorption through that extends over [−2800, −900] km s−1 is present in N V and Lyα. Strong P V absorption isalso detected over −[2200, 1400] km s−1 in this object. The absorption profiles presented here should be taken only asillustrative. They are produced using a simple spline/polynomical fit to the line emission + continuum. The quasartemplate is much too broad to match the observed emission line profiles of Lyα and N V.PG 1435−067.—There are no associated N V or O VI absorption systems in this object, but a faint Lyα absorber
appears to be present at ∼1369.3 Å or ∼ −700 km s−1. The other bluer narrow absorption features are likely Lyα
30 Veilleux et al.
absorption from the intervening CGM.PG 1440+356.—Two blueshifted N V absorbers are observed at −2190 and −1610 km s−1, and a fainter redshifted
one at +460 km s−1 is also detected in both lines of the doublet. Note that the strongest of the blueshifted N V systemsdoes not have a good match in velocity space with Lyα. The strong line near 1334 Å is produced in the Galactic ISM,while the strong line blueward of Lyα is presumably produced by intervening CGM. The Lyα line at +200 km s−1
may be systemic within the uncertainties on the redshift or a signature of inflow.PG 1448+273.—A broad (FWZI ≈ 700 km s−1) multi-component absorption feature is detected in both N V and
Lyα. The narrow Lyα feature at ∼ −2700 km s−1 is likely produced by intervening CGM, while the deep saturatedfeature at 1280 Å is C I from Galactic ISM.PG 1501+106.—Three strong absorption features are seen blueward of Lyα at ∼ 1253.0, 1255.8, and 1257.2 Å,
corresponding to ∼ −1500, −800, and −500 km s−1, respectively (the absorption near 0 km s−1 is at least partly dueto S II from Galactic ISM). Despite the chip gap at the position of the N V doublet, the present data allow us to ruleout the presence of a N V λ1238 counterpart to the most highly blueshifted of these Lyα features since it would lienear ∼1276.5 Å and is not present in the data.PG 1613+658.—Two weak narrow absorption features are detected at −3503 and −3764 km s−1 in N V, O VI,
and Lyα. The other features around N V are of Galactic ISM origin, while those in O VI and Lyα are likely due tointervening CGM.PG 1617+175.—A broad (∼1000 km s−1) blueshifted absorption feature centered around ∼ −3000 km s−1 is detected
in N V, O VI, and Lyα. Three narrow absorbers at −3300, −1630, and −1040 km s−1 are also detected in O VI andLyα but are very weak or absent in N V.PG 1626+554.—The two distinct blueshifted Lyα absorption features at 1378.8 Å and 1374.7 Å, corresponding to−740 and −570 km s−1 in the quasar rest frame, are also detected in Lyβ but not in N V. A faint depression at 1166.9Å may be the −570 km s−1 counterpart of O VI 1032 but it is not detected in the fainter O VI 1038. This feature wasjudged too uncertain to be a detection in our analysis.PG 2130+099.—Two deep and narrow absorption features at ∼ −1500 and 0 km s−1 are detected in N V and Lyα
of this object.PG 2214+139.—Our analysis is based on the co-added spectrum of this object given the lack of variability in the
absorption lines between 2011 and 2012 (Fig. 11). This object shows complex N V troughs that extend from ∼ −3400to −400 km s−1 and are loosely matched to the complex absorption feature at Lyα, except for the sharp Lyα absorptionfeature near −100 km s−1, which is not detected in N V.PG 2233+134.—This object shows a faint and narrow absorber at ∼ −200 km s−1 in O VI and Lyβ. The Lyβ
absorber at ∼ −1300 km s−1 also seems to have a weak O VI counterpart but the fit is inconclusive and is thereforenot included in the statistics for this object. The other stronger lines in this spectral region do not match GalacticISM features so they are likely produced by intervening CGM.PG 2349−014.—There are no N V or O V λ1032 absorbers in this object (O VI λ1038 is lost in the glare of geocoronal
Lyα).
Highly Ionized N V and O VI Outflows in the QUEST Quasars 31
(a)
(b)
(c)
(d
)
(e
)
(f)
(g)
(h)
(
j)
(i)
(k)
(l)
PG00
50+1
24, z
=0.0
59
ï200
0ï1
500
ï100
0ï5
000
500
1000
Velo
city
(km
/s)
0.0
0.2
0.4
0.6
0.8
1.0
Normalized Fh
NV
1238
0.0
0.2
0.4
0.6
0.8
1.0
NV
1242
0.0
0.2
0.4
0.6
0.8
1.0
Ly_
PG08
04+7
61, z
=0.1
00
020
040
060
080
010
00Ve
locit
y (k
m/s
)
0.0
0.2
0.4
0.6
0.8
1.0
Normalized Fh
OVI
103
2
0.0
0.2
0.4
0.6
0.8
1.0
OVI
103
8
0.0
0.2
0.4
0.6
0.8
1.0
NV 1
238
0.0
0.2
0.4
0.6
0.8
1.0
NV 1
242
0.0
0.2
0.4
0.6
0.8
1.0
Ly_
PG08
44+3
49, z
=0.0
64
ï400
ï200
020
040
0Ve
loci
ty (k
m/s
)
0.0
0.2
0.4
0.6
0.8
1.0
Normalized Fh
NV
1238
0.0
0.2
0.4
0.6
0.8
1.0
NV
1242
0.0
0.2
0.4
0.6
0.8
1.0
Ly_
PG09
23+2
01, z
=0.1
92
ï400
0ï3
000
ï200
0ï1
000
0Ve
loci
ty (k
m/s
)
0.0
0.2
0.4
0.6
0.8
1.0
Normalized Fh
OVI
103
2
0.0
0.2
0.4
0.6
0.8
1.0
OVI
103
8
0.0
0.2
0.4
0.6
0.8
1.0
Ly_
0.0
0.2
0.4
0.6
0.8
1.0
Ly`
PG09
53+4
14, z
=0.2
34
ï150
0ï1
000
ï500
050
0Ve
loci
ty (k
m/s
)
0.0
0.2
0.4
0.6
0.8
1.0
Normalized Fh
OVI
103
2
0.0
0.2
0.4
0.6
0.8
1.0
OVI
103
8
0.0
0.2
0.4
0.6
0.8
1.0
Ly`
0.0
0.2
0.4
0.6
0.8
1.0
Lyb
PG10
01+0
54, z
=0.1
61
ï100
00ï8
000
ï600
0ï4
000
ï200
00
Velo
city
(km
/s)
0.0
0.2
0.4
0.6
0.8
1.0
Normalized Fh
OVI
103
2
0.0
0.2
0.4
0.6
0.8
1.0
OVI
103
8
0.0
0.2
0.4
0.6
0.8
1.0
NV
1238
0.0
0.2
0.4
0.6
0.8
1.0
NV
1242
0.0
0.2
0.4
0.6
0.8
1.0
Ly_
0.0
0.2
0.4
0.6
0.8
1.0
Ly`
0.0
0.2
0.4
0.6
0.8
1.0
Lya
PG10
04+1
30, z
=0.2
41
ï1.2×1
04ï9
.0×1
03ï6
.0×1
03ï3
.0×1
030
Velo
city
(km
/s)
0.0
0.2
0.4
0.6
0.8
1.0
Normalized Fh
OVI
103
2
0.0
0.2
0.4
0.6
0.8
1.0
OVI
103
8
0.0
0.2
0.4
0.6
0.8
1.0
Ly`
0.0
0.2
0.4
0.6
0.8
1.0
Lya
PG11
16+2
15, z
=0.1
76
ï400
0ï3
000
ï200
0ï1
000
0Ve
loci
ty (k
m/s
)
0.0
0.2
0.4
0.6
0.8
1.0
Normalized Fh
OVI
103
2
0.0
0.2
0.4
0.6
0.8
1.0
OVI
103
8
0.0
0.2
0.4
0.6
0.8
1.0
Ly_
0.0
0.2
0.4
0.6
0.8
1.0
Ly`
PG11
26ï0
41, z
=0.0
60
ï500
0ï4
000ï3
000ï2
000ï1
000
010
00Ve
loci
ty (k
m/s
)
0.0
0.2
0.4
0.6
0.8
1.0
Normalized Fh
PV 1
117
0.0
0.2
0.4
0.6
0.8
1.0
PV 1
128
0.0
0.2
0.4
0.6
0.8
1.0
OVI
103
2
0.0
0.2
0.4
0.6
0.8
1.0
OVI
103
8
0.0
0.2
0.4
0.6
0.8
1.0
NV
1238
0.0
0.2
0.4
0.6
0.8
1.0
NV
1242
0.0
0.2
0.4
0.6
0.8
1.0
Ly_
0.0
0.2
0.4
0.6
0.8
1.0
Ly`
PG13
07+0
85, z
=0.1
55
ï450
0ï4
000
ï350
0ï3
000
ï250
0Ve
loci
ty (k
m/s
)
0.0
0.2
0.4
0.6
0.8
1.0
Normalized Fh
OVI
103
2
0.0
0.2
0.4
0.6
0.8
1.0
OVI
103
8
0.0
0.2
0.4
0.6
0.8
1.0
Ly_
0.0
0.2
0.4
0.6
0.8
1.0
Ly`
PG13
09+3
55, z
=0.1
83
ï150
0ï1
000
ï500
050
0Ve
loci
ty (k
m/s
)
0.0
0.2
0.4
0.6
0.8
1.0
Normalized Fh
OVI
103
2
0.0
0.2
0.4
0.6
0.8
1.0
OVI
103
8
0.0
0.2
0.4
0.6
0.8
1.0
Ly_
0.0
0.2
0.4
0.6
0.8
1.0
Ly`
0.0
0.2
0.4
0.6
0.8
1.0
Lya
PG13
51+6
40, z
=0.0
88
ï400
0ï3
000
ï200
0ï1
000
0Ve
loci
ty (k
m/s
)
0.0
0.2
0.4
0.6
0.8
1.0
Normalized Fh
NV
1238
0.0
0.2
0.4
0.6
0.8
1.0
NV
1242
0.0
0.2
0.4
0.6
0.8
1.0
Ly_
Fig
ure
14.[PartI]Interlinecompa
risonof
theab
sorbingsystem
sin
each
quasar
withde
tected
NV
orO
VIab
sorption
lines,p
rodu
cedby
plotting
theno
rmalized
spectrum
invelocity
spacerelative
tothequ
asar
rest
fram
e.The
data
arein
black,
theindividu
alcompo
nentsused
tofit
theab
sorption
profi
lesareshow
nin
cyan
,an
dtheoverallfit
isshow
nin
purple.The
velocity
centroidsof
themainab
sorbingsystem
sareindicatedby
vertical
red(L
yα,Lyβ
,N
V,P
V)an
dblue
(OVI)
dotted
lines.
32 Veilleux et al.
(m)
(n)
(o)
(p)
(q
)
(
r)
(s)
(t)
PG14
11+4
42, z
=0.0
90
ï500
0ï4
000
ï300
0ï2
000
ï100
00
1000
Velo
city
(km
/s)
0.0
0.2
0.4
0.6
0.8
1.0
Normalized Fh
PV 1
117
0.0
0.2
0.4
0.6
0.8
1.0
PV 1
128
0.0
0.2
0.4
0.6
0.8
1.0
OVI
103
2
0.0
0.2
0.4
0.6
0.8
1.0
OVI
103
8
0.0
0.2
0.4
0.6
0.8
1.0
NV
1238
0.0
0.2
0.4
0.6
0.8
1.0
NV
1242
0.0
0.2
0.4
0.6
0.8
1.0
Ly_
0.0
0.2
0.4
0.6
0.8
1.0
Ly`
PG14
40+3
56, z
=0.0
77
ï300
0ï2
000
ï100
00
1000
Velo
city
(km
/s)
0.0
0.2
0.4
0.6
0.8
1.0
Normalized Fh
NV
1238
0.0
0.2
0.4
0.6
0.8
1.0
NV
1242
0.0
0.2
0.4
0.6
0.8
1.0
Ly_
PG14
48+2
73, z
=0.0
65
ï600
ï400
ï200
020
0Ve
loci
ty (k
m/s
)
0.0
0.2
0.4
0.6
0.8
1.0
Normalized Fh
NV
1238
0.0
0.2
0.4
0.6
0.8
1.0
NV
1242
0.0
0.2
0.4
0.6
0.8
1.0
Ly_
PG16
13+6
58, z
=0.1
29
ï450
0ï4
000
ï350
0ï3
000
ï250
0Ve
loci
ty (k
m/s
)
0.0
0.2
0.4
0.6
0.8
1.0
Normalized Fh
OVI
103
2
0.0
0.2
0.4
0.6
0.8
1.0
OVI
103
8
0.0
0.2
0.4
0.6
0.8
1.0
NV
1238
0.0
0.2
0.4
0.6
0.8
1.0
NV
1242
0.0
0.2
0.4
0.6
0.8
1.0
Ly_
0.0
0.2
0.4
0.6
0.8
1.0
Ly`
PG16
17+1
75, z
=0.1
14
ï500
0ï4
000
ï300
0ï2
000
ï100
00
Velo
city
(km
/s)
0.0
0.2
0.4
0.6
0.8
1.0
Normalized Fh
OVI
103
2
0.0
0.2
0.4
0.6
0.8
1.0
OVI
103
8
0.0
0.2
0.4
0.6
0.8
1.0
NV
1238
0.0
0.2
0.4
0.6
0.8
1.0
NV
1242
0.0
0.2
0.4
0.6
0.8
1.0
Ly_
0.0
0.2
0.4
0.6
0.8
1.0
Ly`
PG21
30+0
99, z
=0.0
63
ï200
0ï1
500
ï100
0ï5
000
500
Velo
city
(km
/s)
0.0
0.2
0.4
0.6
0.8
1.0
Normalized Fh
NV
1238
0.0
0.2
0.4
0.6
0.8
1.0
NV
1242
0.0
0.2
0.4
0.6
0.8
1.0
Ly_
PG22
14+1
39, z
=0.0
66
ï400
0ï3
000
ï200
0ï1
000
010
00Ve
loci
ty (k
m/s
)
0.0
0.2
0.4
0.6
0.8
1.0
Normalized Fh
NV
1238
0.0
0.2
0.4
0.6
0.8
1.0
NV
1242
0.0
0.2
0.4
0.6
0.8
1.0
Ly_
PG22
33+1
34, z
=0.3
27
ï500
ï400
ï300
ï200
ï100
0Ve
loci
ty (k
m/s
)
0.0
0.2
0.4
0.6
0.8
1.0
Normalized Fh
OVI
103
2
0.0
0.2
0.4
0.6
0.8
1.0
OVI
103
8
0.0
0.2
0.4
0.6
0.8
1.0
Ly`
Fig
ure
13.[PartII]S
ameas
previous
figure.
Highly Ionized N V and O VI Outflows in the QUEST Quasars 33
REFERENCES
Allen, J. T., Hewett, P. C., Maddox, N., Richards, G. T., &Belokurov, V. 2011, MNRAS, 410, 860,doi: 10.1111/j.1365-2966.2010.17489.x
Arav, N., Kaastra, J., Kriss, G. A., et al. 2005, ApJ, 620,665, doi: 10.1086/425560
Arav, N., & Li, Z.-Y. 1994, ApJ, 427, 700,doi: 10.1086/174177
Arav, N., Li, Z.-Y., & Begelman, M. C. 1994, ApJ, 432, 62,doi: 10.1086/174549
Arav, N., Moe, M., Costantini, E., et al. 2008, ApJ, 681,954, doi: 10.1086/588651
Arav, N., Xu, X., Miller, T., Kriss, G. A., & Plesha, R.2020, ApJS, 247, 37, doi: 10.3847/1538-4365/ab66af
Arav, N., Chamberlain, C., Kriss, G. A., et al. 2015, A&A,577, A37, doi: 10.1051/0004-6361/201425302
Banda-Barragán, W. E., Brüggen, M., Federrath, C., et al.2020, MNRAS, 499, 2173, doi: 10.1093/mnras/staa2904
Banda-Barragán, W. E., Federrath, C., Crocker, R. M., &Bicknell, G. V. 2018, MNRAS, 473, 3454,doi: 10.1093/mnras/stx2541
Banda-Barragán, W. E., Parkin, E. R., Federrath, C.,Crocker, R. M., & Bicknell, G. V. 2016, MNRAS, 455,1309, doi: 10.1093/mnras/stv2405
Barvainis, R. 1987, ApJ, 320, 537, doi: 10.1086/165571Baskin, A., & Laor, A. 2005, MNRAS, 356, 1029,doi: 10.1111/j.1365-2966.2004.08525.x
—. 2018, MNRAS, 474, 1970, doi: 10.1093/mnras/stx2850Baskin, A., Laor, A., & Hamann, F. 2013, MNRAS, 432,1525, doi: 10.1093/mnras/stt582
Baskin, A., Laor, A., & Stern, J. 2014, MNRAS, 445, 3025,doi: 10.1093/mnras/stu1732
Becker, R. H., White, R. L., Gregg, M. D., et al. 2001,ApJS, 135, 227, doi: 10.1086/321798
Benn, C. R., Carballo, R., Holt, J., et al. 2005, MNRAS,360, 1455, doi: 10.1111/j.1365-2966.2005.09143.x
Bentz, M. C., & Katz, S. 2015, PASP, 127, 67,doi: 10.1086/679601
Bentz, M. C., Denney, K. D., Grier, C. J., et al. 2013, ApJ,767, 149, doi: 10.1088/0004-637X/767/2/149
Borguet, B. C. J., Arav, N., Edmonds, D., Chamberlain,C., & Benn, C. 2013, ApJ, 762, 49,doi: 10.1088/0004-637X/762/1/49
Borguet, B. C. J., Edmonds, D., Arav, N., Benn, C., &Chamberlain, C. 2012, ApJ, 758, 69,doi: 10.1088/0004-637X/758/1/69
Boroson, T. A., & Green, R. F. 1992, ApJS, 80, 109,doi: 10.1086/191661
Bowler, R. A. A., Hewett, P. C., Allen, J. T., & Ferland,G. J. 2014, MNRAS, 445, 359,doi: 10.1093/mnras/stu1730
Brandt, W. N., Laor, A., & Wills, B. J. 2000, ApJ, 528,637, doi: 10.1086/308207
Braun, E., & Milgrom, M. 1989, ApJ, 342, 100,doi: 10.1086/167578
Cameron, E. 2011, PASA, 28, 128, doi: 10.1071/AS10046Capellupo, D. M., Hamann, F., Shields, J. C., RodríguezHidalgo, P., & Barlow, T. A. 2012, MNRAS, 422, 3249,doi: 10.1111/j.1365-2966.2012.20846.x
Capellupo, D. M., Hamann, F., Herbst, H., et al. 2017,MNRAS, 469, 323, doi: 10.1093/mnras/stx870
Carilli, C. L., Wrobel, J. M., & Ulvestad, J. S. 1998, AJ,115, 928, doi: 10.1086/300253
Cicone, C., Maiolino, R., Sturm, E., et al. 2014, A&A, 562,A21, doi: 10.1051/0004-6361/201322464
Cooksey, K. L., Prochaska, J. X., Chen, H.-W., Mulchaey,J. S., & Weiner, B. J. 2008, ApJ, 676, 262,doi: 10.1086/528704
Cooper, J. L., Bicknell, G. V., Sutherland, R. S., &Bland-Hawthorn, J. 2009, ApJ, 703, 330,doi: 10.1088/0004-637X/703/1/330
Crenshaw, D. M., & Kraemer, S. B. 2012, ApJ, 753, 75,doi: 10.1088/0004-637X/753/1/75
Crenshaw, D. M., Kraemer, S. B., Boggess, A., et al. 1999,ApJ, 516, 750, doi: 10.1086/307144
Curran, P. A. 2014, arXiv e-prints, arXiv:1411.3816.https://arxiv.org/abs/1411.3816
Danehkar, A., Nowak, M. A., Lee, J. C., et al. 2018, ApJ,853, 165, doi: 10.3847/1538-4357/aaa427
Danforth, C. W., Keeney, B. A., Stocke, J. T., Shull, J. M.,& Yao, Y. 2010, ApJ, 720, 976,doi: 10.1088/0004-637X/720/1/976
Danforth, C. W., Keeney, B. A., Tilton, E. M., et al. 2016,ApJ, 817, 111, doi: 10.3847/0004-637X/817/2/111
Dasyra, K. M., Tacconi, L. J., Davies, R. I., et al. 2007,ApJ, 657, 102, doi: 10.1086/510552
de Vaucouleurs, G., de Vaucouleurs, A., Corwin, Herold G.,J., et al. 1991, Third Reference Catalogue of BrightGalaxies
Dunn, J. P., Crenshaw, D. M., Kraemer, S. B., & Trippe,M. L. 2008, AJ, 136, 1201,doi: 10.1088/0004-6256/136/3/1201
Efstathiou, G. 2000, MNRAS, 317, 697,doi: 10.1046/j.1365-8711.2000.03665.x
Elvis, M. 2000, ApJ, 545, 63, doi: 10.1086/317778Evans, A. S., Solomon, P. M., Tacconi, L. J., Vavilkin, T.,& Downes, D. 2006, AJ, 132, 2398, doi: 10.1086/508416
34 Veilleux et al.
Faucher-Giguère, C.-A., & Quataert, E. 2012, MNRAS,425, 605, doi: 10.1111/j.1365-2966.2012.21512.x
Filiz Ak, N., Brandt, W. N., Hall, P. B., et al. 2012, ApJ,757, 114, doi: 10.1088/0004-637X/757/2/114
—. 2013, ApJ, 777, 168, doi: 10.1088/0004-637X/777/2/168Fitzpatrick, E. L. 1999, PASP, 111, 63, doi: 10.1086/316293Fluetsch, A., Maiolino, R., Carniani, S., et al. 2019,MNRAS, 483, 4586, doi: 10.1093/mnras/sty3449
—. 2020, arXiv e-prints, arXiv:2006.13232.https://arxiv.org/abs/2006.13232
Foltz, C. B., Weymann, R. J., Morris, S. L., & Turnshek,D. A. 1987, ApJ, 317, 450, doi: 10.1086/165290
Gabel, J. R., Crenshaw, D. M., Kraemer, S. B., et al. 2003,ApJ, 583, 178, doi: 10.1086/345096
Ganguly, R., Bond, N. A., Charlton, J. C., et al. 2001, ApJ,549, 133, doi: 10.1086/319082
Ganguly, R., & Brotherton, M. S. 2008, ApJ, 672, 102,doi: 10.1086/524106
Ganguly, R., Brotherton, M. S., Cales, S., et al. 2007, ApJ,665, 990, doi: 10.1086/519759
Ganguly, R., Masiero, J., Charlton, J. C., & Sembach,K. R. 2003, ApJ, 598, 922, doi: 10.1086/379057
George, I. M., Turner, T. J., Netzer, H., et al. 1998, ApJS,114, 73, doi: 10.1086/313067
Gibson, R. R., Brandt, W. N., Gallagher, S. C., Hewett,P. C., & Schneider, D. P. 2010, ApJ, 713, 220,doi: 10.1088/0004-637X/713/1/220
Gibson, R. R., Brandt, W. N., Gallagher, S. C., &Schneider, D. P. 2009a, ApJ, 696, 924,doi: 10.1088/0004-637X/696/1/924
Gibson, R. R., Brandt, W. N., Schneider, D. P., &Gallagher, S. C. 2008, ApJ, 675, 985, doi: 10.1086/527462
Gibson, R. R., Jiang, L., Brandt, W. N., et al. 2009b, ApJ,692, 758, doi: 10.1088/0004-637X/692/1/758
Girichidis, P., Naab, T., Walch, S., & Berlok, T. 2021,MNRAS, doi: 10.1093/mnras/stab1203
Giustini, M., & Proga, D. 2019, A&A, 630, A94,doi: 10.1051/0004-6361/201833810
Giustini, M., Cappi, M., Chartas, G., et al. 2011, A&A,536, A49, doi: 10.1051/0004-6361/201117732
Gravity Collaboration, Sturm, E., Dexter, J., et al. 2018,Nature, 563, 657, doi: 10.1038/s41586-018-0731-9
Grier, C. J., Hall, P. B., Brandt, W. N., et al. 2015, ApJ,806, 111, doi: 10.1088/0004-637X/806/1/111
Gronke, M., & Oh, S. P. 2018, MNRAS, 480, L111,doi: 10.1093/mnrasl/sly131
—. 2020, MNRAS, 492, 1970, doi: 10.1093/mnras/stz3332Grønnow, A., Tepper-García, T., & Bland-Hawthorn, J.2018, ApJ, 865, 64, doi: 10.3847/1538-4357/aada0e
Guyon, O., Sanders, D. B., & Stockton, A. 2006, ApJS,166, 89, doi: 10.1086/505030
Haislmaier, K. J., Tripp, T. M., Katz, N., et al. 2021,MNRAS, 502, 4993, doi: 10.1093/mnras/staa3544
Hamann, F. 1998, ApJ, 500, 798, doi: 10.1086/305776Hamann, F., Barlow, T. A., Junkkarinen, V., & Burbidge,E. M. 1997a, ApJ, 478, 80, doi: 10.1086/303781
Hamann, F., Beaver, E. A., Cohen, R. D., et al. 1997b,ApJ, 488, 155, doi: 10.1086/304698
Hamann, F., Chartas, G., McGraw, S., et al. 2013,MNRAS, 435, 133, doi: 10.1093/mnras/stt1231
Hamann, F., Herbst, H., Paris, I., & Capellupo, D. 2019a,MNRAS, 483, 1808, doi: 10.1093/mnras/sty2900
Hamann, F., Kanekar, N., Prochaska, J. X., et al. 2011,MNRAS, 410, 1957,doi: 10.1111/j.1365-2966.2010.17575.x
Hamann, F., Kaplan, K. F., Rodríguez Hidalgo, P.,Prochaska, J. X., & Herbert-Fort, S. 2008, MNRAS, 391,L39, doi: 10.1111/j.1745-3933.2008.00554.x
Hamann, F., & Sabra, B. 2004, in Astronomical Society ofthe Pacific Conference Series, Vol. 311, AGN Physicswith the Sloan Digital Sky Survey, ed. G. T. Richards &P. B. Hall, 203. https://arxiv.org/abs/astro-ph/0310668
Hamann, F., Tripp, T. M., Rupke, D., & Veilleux, S. 2019b,MNRAS, 487, 5041, doi: 10.1093/mnras/stz1408
Hamilton, T. S., Casertano, S., & Turnshek, D. A. 2008,ApJ, 678, 22, doi: 10.1086/524182
Harris, D. W., Jensen, T. W., Suzuki, N., et al. 2016, AJ,151, 155, doi: 10.3847/0004-6256/151/6/155
He, Z., Wang, T., Liu, G., et al. 2019, Nature Astronomy,3, 265, doi: 10.1038/s41550-018-0669-8
Hewett, P. C., & Wild, V. 2010, MNRAS, 405, 2302,doi: 10.1111/j.1365-2966.2010.16648.x
Hickox, R. C., & Alexander, D. M. 2018, ARA&A, 56, 625,doi: 10.1146/annurev-astro-081817-051803
Ho, L. C., & Kim, M. 2009, ApJS, 184, 398,doi: 10.1088/0067-0049/184/2/398
Hopkins, P. F., & Elvis, M. 2010, MNRAS, 401, 7,doi: 10.1111/j.1365-2966.2009.15643.x
Hopkins, P. F., Grudić, M. Y., Wetzel, A., et al. 2020,MNRAS, 491, 3702, doi: 10.1093/mnras/stz3129
Hopkins, P. F., Kereš, D., Oñorbe, J., et al. 2014, MNRAS,445, 581, doi: 10.1093/mnras/stu1738
Hu, C., Li, S.-S., Guo, W.-J., et al. 2020, ApJ, 905, 75,doi: 10.3847/1538-4357/abc2da
Inoue, H., Terashima, Y., & Ho, L. C. 2007, ApJ, 662, 860,doi: 10.1086/517995
Ishibashi, W., Fabian, A. C., & Arakawa, N. 2021,MNRAS, 502, 3638, doi: 10.1093/mnras/stab266
Highly Ionized N V and O VI Outflows in the QUEST Quasars 35
Ishibashi, W., Fabian, A. C., & Maiolino, R. 2018,MNRAS, 476, 512, doi: 10.1093/mnras/sty236
Isobe, T., Feigelson, E. D., & Nelson, P. I. 1986, ApJ, 306,490, doi: 10.1086/164359
Jin, C., Ward, M., Done, C., & Gelbord, J. 2012, MNRAS,420, 1825, doi: 10.1111/j.1365-2966.2011.19805.x
Kaspi, S., Brandt, W. N., Maoz, D., et al. 2007, ApJ, 659,997, doi: 10.1086/512094
Kaspi, S., Brandt, W. N., Netzer, H., et al. 2000, ApJL,535, L17, doi: 10.1086/312697
Kaspi, S., Maoz, D., Netzer, H., et al. 2005, ApJ, 629, 61,doi: 10.1086/431275
Kaspi, S., Brandt, W. N., Netzer, H., et al. 2001, ApJ, 554,216, doi: 10.1086/321333
Kelly, B. C. 2007, ApJ, 665, 1489, doi: 10.1086/519947Kim, M., Ho, L. C., Peng, C. Y., Barth, A. J., & Im, M.2008, ApJS, 179, 283, doi: 10.1086/591796
Knigge, C., Scaringi, S., Goad, M. R., & Cottis, C. E. 2008,MNRAS, 386, 1426,doi: 10.1111/j.1365-2966.2008.13081.x
Kraemer, S. B., Crenshaw, D. M., & Gabel, J. R. 2001,ApJ, 557, 30, doi: 10.1086/321619
Kriss, G. A. 2004a, arXiv e-prints, astro.https://arxiv.org/abs/astro-ph/0411380
Kriss, G. A. 2004b, in The Interplay Among Black Holes,Stars and ISM in Galactic Nuclei, ed.T. Storchi-Bergmann, L. C. Ho, & H. R. Schmitt, Vol.222, 223–228, doi: 10.1017/S1743921304002157
Kriss, G. A., Lee, J. C., Danehkar, A., et al. 2018, ApJ,853, 166, doi: 10.3847/1538-4357/aaa42b
Krongold, Y., Nicastro, F., Brickhouse, N. S., et al. 2003,ApJ, 597, 832, doi: 10.1086/378639
Krug, H. B. 2013, PhD thesis, University of Maryland,College Park
Laha, S., Guainazzi, M., Piconcelli, E., et al. 2018, ApJ,868, 10, doi: 10.3847/1538-4357/aae390
Laha, S., Reynolds, C. S., Reeves, J., et al. 2021, NatureAstronomy, 5, 13, doi: 10.1038/s41550-020-01255-2
Lani, C., Netzer, H., & Lutz, D. 2017, MNRAS, 471, 59,doi: 10.1093/mnras/stx1374
Laor, A., & Brandt, W. N. 2002, ApJ, 569, 641,doi: 10.1086/339476
Laor, A., & Davis, S. W. 2014, MNRAS, 438, 3024,doi: 10.1093/mnras/stt2408
Leighly, K. M., Dietrich, M., & Barber, S. 2011, ApJ, 728,94, doi: 10.1088/0004-637X/728/2/94
Lodders, K. 2003, ApJ, 591, 1220, doi: 10.1086/375492Lu, W.-J., & Lin, Y.-R. 2018, ApJ, 863, 186,doi: 10.3847/1538-4357/aad411
Luo, B., Brandt, W. N., Hall, P. B., et al. 2015, ApJ, 805,122, doi: 10.1088/0004-637X/805/2/122
Lutz, D., Sturm, E., Janssen, A., et al. 2020, A&A, 633,A134, doi: 10.1051/0004-6361/201936803
Lyu, J., Rieke, G. H., & Shi, Y. 2017, ApJ, 835, 257,doi: 10.3847/1538-4357/835/2/257
Marcolini, A., Strickland, D. K., D’Ercole, A., Heckman,T. M., & Hoopes, C. G. 2005, MNRAS, 362, 626,doi: 10.1111/j.1365-2966.2005.09343.x
Martin, C. L., Dijkstra, M., Henry, A., et al. 2015, ApJ,803, 6, doi: 10.1088/0004-637X/803/1/6
Marziani, P., Sulentic, J. W., Dultzin-Hacyan, D., Calvani,M., & Moles, M. 1996, ApJS, 104, 37,doi: 10.1086/192291
Mas-Ribas, L., & Mauland, R. 2019, ApJ, 886, 151,doi: 10.3847/1538-4357/ab4efd
Moravec, E. A., Hamann, F., Capellupo, D. M., et al. 2017,MNRAS, 468, 4539, doi: 10.1093/mnras/stx775
Murray, N., Chiang, J., Grossman, S. A., & Voit, G. M.1995, ApJ, 451, 498, doi: 10.1086/176238
Mushotzky, R. F., Solomon, P. M., & Strittmatter, P. A.1972, ApJ, 174, 7, doi: 10.1086/151463
Nelson, D., Pillepich, A., Springel, V., et al. 2019, MNRAS,490, 3234, doi: 10.1093/mnras/stz2306
Nelson, E. J., Tacchella, S., Diemer, B., et al. 2021, arXive-prints, arXiv:2101.12212.https://arxiv.org/abs/2101.12212
Netzer, H., & Laor, A. 1993, ApJL, 404, L51,doi: 10.1086/186741
Netzer, H., Lutz, D., Schweitzer, M., et al. 2007, ApJ, 666,806, doi: 10.1086/520716
Nims, J., Quataert, E., & Faucher-Giguère, C.-A. 2015,MNRAS, 447, 3612, doi: 10.1093/mnras/stu2648
Onken, C. A., Ferrarese, L., Merritt, D., et al. 2004, ApJ,615, 645, doi: 10.1086/424655
Oppenheimer, B. D., Davies, J. J., Crain, R. A., et al. 2020,MNRAS, 491, 2939, doi: 10.1093/mnras/stz3124
Perry, J. J., & O’Dell, S. L. 1978, A&A, 62, 229Piconcelli, E., Jimenez-Bailón, E., Guainazzi, M., et al.2005, A&A, 432, 15, doi: 10.1051/0004-6361:20041621
Privon, G. C., Ricci, C., Aalto, S., et al. 2020, ApJ, 893,149, doi: 10.3847/1538-4357/ab8015
Prochaska, J. X., Wolfe, A. M., Tytler, D., et al. 2001,ApJS, 137, 21, doi: 10.1086/322542
Proga, D. 2007, ApJ, 661, 693, doi: 10.1086/515389Proga, D., & Kallman, T. R. 2004, ApJ, 616, 688,doi: 10.1086/425117
Rankine, A. L., Hewett, P. C., Banerji, M., & Richards,G. T. 2020, MNRAS, 492, 4553,doi: 10.1093/mnras/staa130
36 Veilleux et al.
Revalski, M., Crenshaw, D. M., Kraemer, S. B., et al. 2018,ApJ, 856, 46, doi: 10.3847/1538-4357/aab107
Reynolds, C. S. 1997, MNRAS, 286, 513,doi: 10.1093/mnras/286.3.513
Ricci, C., Trakhtenbrot, B., Koss, M. J., et al. 2017, ApJS,233, 17, doi: 10.3847/1538-4365/aa96ad
Richards, G. T., Kruczek, N. E., Gallagher, S. C., et al.2011, AJ, 141, 167, doi: 10.1088/0004-6256/141/5/167
Richings, A. J., & Faucher-Giguère, C.-A. 2018a, MNRAS,474, 3673, doi: 10.1093/mnras/stx3014
—. 2018b, MNRAS, 478, 3100, doi: 10.1093/mnras/sty1285Richings, A. J., Faucher-Giguère, C.-A., & Stern, J. 2021,MNRAS, 503, 1568, doi: 10.1093/mnras/stab556
Rupke, D. 2021a, doi: 10.5281/zenodo.5659382—. 2021b, doi: 10.5281/zenodo.5659520—. 2021c, doi: 10.5281/zenodo.5659554Rupke, D. S., Veilleux, S., & Sanders, D. B. 2005, ApJS,160, 87, doi: 10.1086/432886
Rupke, D. S. N. 2014, IFSFIT: Spectral Fitting for IntegralField Spectrographs. http://ascl.net/1409.005
Rupke, D. S. N., Gültekin, K., & Veilleux, S. 2017, ApJ,850, 40, doi: 10.3847/1538-4357/aa94d1
Rupke, D. S. N., & Veilleux, S. 2015, ApJ, 801, 126,doi: 10.1088/0004-637X/801/2/126
Sanders, D. B., Soifer, B. T., Elias, J. H., et al. 1988, ApJ,325, 74, doi: 10.1086/165983
Savage, B. D., Kim, T. S., Wakker, B. P., et al. 2014, ApJS,212, 8, doi: 10.1088/0067-0049/212/1/8
Scannapieco, E., & Brüggen, M. 2015, ApJ, 805, 158,doi: 10.1088/0004-637X/805/2/158
Scargle, J. D. 1973, ApJ, 179, 705, doi: 10.1086/151910Schlafly, E. F., & Finkbeiner, D. P. 2011, ApJ, 737, 103,doi: 10.1088/0004-637X/737/2/103
Schmidt, M., & Green, R. F. 1983, ApJ, 269, 352,doi: 10.1086/161048
Schneider, D. P., Richards, G. T., Hall, P. B., et al. 2010,AJ, 139, 2360, doi: 10.1088/0004-6256/139/6/2360
Schneider, E. E., Ostriker, E. C., Robertson, B. E., &Thompson, T. A. 2020, ApJ, 895, 43,doi: 10.3847/1538-4357/ab8ae8
Schneider, E. E., & Robertson, B. E. 2015, ApJS, 217, 24,doi: 10.1088/0067-0049/217/2/24
—. 2017, ApJ, 834, 144, doi: 10.3847/1538-4357/834/2/144Schweitzer, M., Lutz, D., Sturm, E., et al. 2006, ApJ, 649,79, doi: 10.1086/506510
Schweitzer, M., Groves, B., Netzer, H., et al. 2008, ApJ,679, 101, doi: 10.1086/587097
Shangguan, J., Ho, L. C., & Xie, Y. 2018, ApJ, 854, 158,doi: 10.3847/1538-4357/aaa9be
Shankar, F., Dai, X., & Sivakoff, G. R. 2008, ApJ, 687, 859,doi: 10.1086/591488
Silich, S., Tenorio-Tagle, G., & Muñoz-Tuñón, C. 2003,ApJ, 590, 791, doi: 10.1086/375133
Silva, C. V., Costantini, E., Giustini, M., et al. 2018,MNRAS, 480, 2334, doi: 10.1093/mnras/sty1938
Sim, S. A., Proga, D., Miller, L., Long, K. S., & Turner,T. J. 2010, MNRAS, 408, 1396,doi: 10.1111/j.1365-2966.2010.17215.x
Somalwar, J., Johnson, S. D., Stern, J., et al. 2020, ApJL,890, L28, doi: 10.3847/2041-8213/ab733d
Springob, C. M., Haynes, M. P., Giovanelli, R., & Kent,B. R. 2005, ApJS, 160, 149, doi: 10.1086/431550
Srianand, R. 2000, ApJ, 528, 617, doi: 10.1086/308214Srianand, R., Petitjean, P., Ledoux, C., & Hazard, C. 2002,MNRAS, 336, 753, doi: 10.1046/j.1365-8711.2002.05792.x
Stern, J., Faucher-Giguère, C.-A., Zakamska, N. L., &Hennawi, J. F. 2016, ApJ, 819, 130,doi: 10.3847/0004-637X/819/2/130
Stevans, M. L., Shull, J. M., Danforth, C. W., & Tilton,E. M. 2014, ApJ, 794, 75,doi: 10.1088/0004-637X/794/1/75
Sturm, E., González-Alfonso, E., Veilleux, S., et al. 2011,ApJL, 733, L16, doi: 10.1088/2041-8205/733/1/L16
Surace, J. A., Sanders, D. B., & Evans, A. S. 2001, AJ, 122,2791, doi: 10.1086/324462
Teng, S. H., & Veilleux, S. 2010, ApJ, 725, 1848,doi: 10.1088/0004-637X/725/2/1848
Teng, S. H., Veilleux, S., & Baker, A. J. 2013, ApJ, 765, 95,doi: 10.1088/0004-637X/765/2/95
Teng, S. H., Brandt, W. N., Harrison, F. A., et al. 2014,ApJ, 785, 19, doi: 10.1088/0004-637X/785/1/19
Thompson, T. A., Quataert, E., Zhang, D., & Weinberg,D. H. 2016, MNRAS, 455, 1830,doi: 10.1093/mnras/stv2428
Tombesi, F., Meléndez, M., Veilleux, S., et al. 2015, Nature,519, 436, doi: 10.1038/nature14261
Tombesi, F., Veilleux, S., Meléndez, M., et al. 2017, ApJ,850, 151, doi: 10.3847/1538-4357/aa9579
Tripp, T. M., Lu, L., & Savage, B. D. 1998, ApJ, 508, 200,doi: 10.1086/306397
Tripp, T. M., Sembach, K. R., Bowen, D. V., et al. 2008,ApJS, 177, 39, doi: 10.1086/587486
Trump, J. R., Hall, P. B., Reichard, T. A., et al. 2006,ApJS, 165, 1, doi: 10.1086/503834
U, V., Medling, A., Sanders, D., et al. 2013, ApJ, 775, 115,doi: 10.1088/0004-637X/775/2/115
Vanden Berk, D. E., Richards, G. T., Bauer, A., et al. 2001,AJ, 122, 549, doi: 10.1086/321167
Highly Ionized N V and O VI Outflows in the QUEST Quasars 37
Veilleux, S., Bolatto, A., Tombesi, F., et al. 2017, ApJ, 843,18, doi: 10.3847/1538-4357/aa767d
Veilleux, S., Kim, D. C., & Sanders, D. B. 2002, ApJS, 143,315, doi: 10.1086/343844
Veilleux, S., Maiolino, R., Bolatto, A. D., & Aalto, S. 2020,A&A Rv, 28, 2, doi: 10.1007/s00159-019-0121-9
Veilleux, S., Meléndez, M., Tripp, T. M., Hamann, F., &Rupke, D. S. N. 2016, ApJ, 825, 42,doi: 10.3847/0004-637X/825/1/42
Veilleux, S., Teng, S. H., Rupke, D. S. N., Maiolino, R., &Sturm, E. 2014, ApJ, 790, 116,doi: 10.1088/0004-637X/790/2/116
Veilleux, S., Kim, D. C., Peng, C. Y., et al. 2006, ApJ, 643,707, doi: 10.1086/503188
Veilleux, S., Rupke, D. S. N., Kim, D. C., et al. 2009a,ApJS, 182, 628, doi: 10.1088/0067-0049/182/2/628
Veilleux, S., Kim, D. C., Rupke, D. S. N., et al. 2009b, ApJ,701, 587, doi: 10.1088/0004-637X/701/1/587
Veilleux, S., Trippe, M., Hamann, F., et al. 2013a, ApJ,764, 15, doi: 10.1088/0004-637X/764/1/15
Veilleux, S., Meléndez, M., Sturm, E., et al. 2013b, ApJ,776, 27, doi: 10.1088/0004-637X/776/1/27
Vestergaard, M., & Peterson, B. M. 2006, ApJ, 641, 689,doi: 10.1086/500572
Virtanen, P., Gommers, R., Oliphant, T. E., et al. 2020,Nature Methods, 17, 261, doi: 10.1038/s41592-019-0686-2
Waddell, S. G. H., & Gallo, L. C. 2020, MNRAS, 498, 5207,doi: 10.1093/mnras/staa2783
Wang, T., & Wang, J. 1999, Publications of the PurpleMountain Observatory, 18, 248
Wang, T. G., Brinkmann, W., Wamsteker, W., Yuan, W.,& Wang, J. X. 1999, MNRAS, 307, 821,doi: 10.1046/j.1365-8711.1999.02650.x
Werk, J. K., Prochaska, J. X., Cantalupo, S., et al. 2016,ApJ, 833, 54, doi: 10.3847/1538-4357/833/1/54
Weymann, R. J., Carswell, R. F., & Smith, M. G. 1981,ARA&A, 19, 41,doi: 10.1146/annurev.aa.19.090181.000353
Weymann, R. J., Morris, S. L., Foltz, C. B., & Hewett,P. C. 1991, ApJ, 373, 23, doi: 10.1086/170020
Weymann, R. J., Turnshek, D. A., & Christiansen, W. A.1985, in Astrophysics of Active Galaxies andQuasi-Stellar Objects, ed. J. S. Miller, 333–365
Wills, B. J., Brandt, W. N., & Laor, A. 1999, ApJL, 520,L91, doi: 10.1086/312165
Wu, J., Brandt, W. N., Hall, P. B., et al. 2011, ApJ, 736,28, doi: 10.1088/0004-637X/736/1/28
Yi, W., Brandt, W. N., Hall, P. B., et al. 2019, ApJS, 242,28, doi: 10.3847/1538-4365/ab1f90
Zhang, S., Wang, H., Wang, T., et al. 2014, ApJ, 786, 42,doi: 10.1088/0004-637X/786/1/42
Zhuang, M.-Y., Ho, L. C., & Shangguan, J. 2018, ApJ, 862,118, doi: 10.3847/1538-4357/aacc2d
Zubovas, K., & King, A. 2012, ApJL, 745, L34,doi: 10.1088/2041-8205/745/2/L34
Zubovas, K., & King, A. R. 2014, MNRAS, 439, 400,doi: 10.1093/mnras/stt2472